EVOLUTION OF PHOTOSYMBIOSIS IN MARINE BIVALVES (CARDIIDAE: FRAGINAE) By LISA ANN KIRKENDALE A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2005
Copyright 2005 by Lisa Ann Kirkendale
This document is dedicated to my gr andfather, the late John Kirkendale.
iv ACKNOWLEDGMENTS I would like to thank the staff, faculty a nd graduate students in the Department of Zoology and the Florida Museum of Natural Hist ory at the University of Florida for their support and advice. My committee and especial ly my advisor Gustav Paulay were very helpful at all stages of my doctoral res earch. The study would not have been possible without funding provided from a number of sources including Lerner Gray Fund for Marine Research, Sigma Xi Grants in Ai d of Research, Astronaut Trail Shell Club, Western Society of Malacologists and Conchol ogists of America. I would also like to mention the Paris Museum of Natural Hist ory for an opportunity to work at that institution. My family has offered immense em otional (not to men tion financial) support through the years. I am grateful to Peter Middelfart for his love and understanding.
v TABLE OF CONTENTS page ACKNOWLEDGMENTS.................................................................................................iv LIST OF TABLES...........................................................................................................viii LIST OF FIGURES...........................................................................................................ix ABSTRACT....................................................................................................................... xi CHAPTER 1 GENERAL INTRODUCTION....................................................................................1 2 MOLECULAR PHYLOGENETICS..........................................................................10 Introduction.................................................................................................................10 Higher-Level Systematics of the Cardioidea and Cardiidae...............................13 Fraginae Monophyly...........................................................................................15 Parvicardium and Papillicardium : Sister to all other Fragines?.........................16 The Â“ Trigoniocardia Â” and Â“ Ctenocardia Â” groups...............................................19 Known Photosymbiotic Representatives: Fragum, Lunulicardia and Corculum ..........................................................................................................20 Materials and Methods...............................................................................................24 Specimen Acquisition..........................................................................................24 DNA Extraction and PCR...................................................................................24 Alignment & Molecular Analyses.......................................................................27 Results........................................................................................................................ .29 General FindingsSequence Data.......................................................................29 General FindingsPhylogenetic Methods...........................................................29 Higher-Level Phylogenetic Findings...................................................................30 Fraginae Monophyly...........................................................................................31 The Genus Parvicardium sensu lato: Clades I and IV.......................................35 The Â“ CtenocardiaÂ” and Â“ TrigoniocardiaÂ” groups: Clade II..............................35 Fragum , Lunulicardia and Corculum : Clade III................................................36 SpeciesÂ’ Boundaries in IWP Corculum , Lunulicardia and Fragum ...................36 Discussion...................................................................................................................39 Monophyly of the Cardiid Subfamily Fraginae?.................................................39 The Early Diverging Fragine genus Parvicardium : Clades I and IV.................41 The Â“ CtenocardiaÂ” and Â“ TrigoniocardiaÂ” groups: Clade II..............................43
vi Relationships among the IWP genera Fragum , Corculum and Lunulicardia : Clade III...........................................................................................................44 3 CHARACTER TRAIT EVOLUTION.......................................................................49 Introduction.................................................................................................................49 Materials and Methods...............................................................................................56 Specimen Acquisition..........................................................................................56 Determination of Photosymbiotic Condition.......................................................56 Morphological Character Trait Analyses............................................................58 Character Data.....................................................................................................58 Windows.......................................................................................................58 Maximum observed adult size......................................................................61 Habitat..........................................................................................................62 Shell measurements......................................................................................64 Character Trait Evolution....................................................................................66 Results........................................................................................................................ .66 Photosymbiotic Status.........................................................................................66 Character Trait States..........................................................................................67 Windows.......................................................................................................67 Maximum observed adult size......................................................................67 Habitat..........................................................................................................69 Shell measurements......................................................................................70 Character Trait Evolution....................................................................................74 Windows.......................................................................................................77 Maximum observed adult size......................................................................78 Habitat..........................................................................................................81 Shell measurements......................................................................................84 Discussion...................................................................................................................89 Photosymbiotic Status.........................................................................................89 Character Trait Analyses and Evolution..............................................................89 Windows.......................................................................................................89 Maximum observed adult size......................................................................90 Habitat..........................................................................................................91 Shell measurements......................................................................................91 4 GLOBAL DISCUSSION............................................................................................94 Molecular phylogenetics.............................................................................................94 Character trait evolution.............................................................................................97 APPENDIX A NUMBER OF INDIVIDUALS EXAMINED FOR SCORED CHARACTERS OF FRAGINAE..............................................................................................................101
vii B PHOTOSYMBIOTIC STATUS, MORPHOMETRIC AND HABITAT DATA FOR MEASURED CHARACTERS........................................................................103 C RAW LOGGED DATA FOR 95% C ONFIDENCE INTERVAL PLOTS OF KEEL ANGLE, POSTERIOR SHELL FLATTENING AND LINEARIZED SURFACE AREA TO VOLUME RATIOS.............................................................105 LIST OF REFERENCES.................................................................................................107 BIOGRAPHICAL SKETCH...........................................................................................118
viii LIST OF TABLES Table page 2-1 Membership in the subfamily Fraginae....................................................................16 2-2 Species and distributi on of recognized Fraginae......................................................18 2-3 Sampling localities and voucher information for a subset of Fraginae....................25 2-4 Primers used for phylogenetic recons tructions of Fraginae relationships................26 3-1 Species, distribution and vouc hers of sampled Fraginae.........................................57 3-2 Photosymbiotic status of sampled Fraginae.............................................................68 3-3 Morphological character matrix for mapping traits of sampled Fraginae................69
ix LIST OF FIGURES Figure page 2-1 Schematic diagram of some shell adaptations.......................................................11 2-2 Recent cardiid and fragine phylogenies.................................................................14 2-3 Overview of major clades and support..................................................................31 2-4 Bayesian 50% majority topology for the 28S gene region....................................32 2-5 Bayesian 50% majority topol ogy for the mitochondrial dataset............................33 2-6 Bayesian 50% majority topology for the mitochondrial and 28S datasets............34 2-7 COI phylogram (NJ, GTR model) of speciesÂ’ boundaries and substructuring......38 2-8 COI phylogram (NJ, GTR mode l) of speciesÂ’ boundaries in Corculum. ..............39 2-9 Geographic signature of major Fraginae groups....................................................46 3-1 Schematic diagram of some shell adaptations.......................................................54 3-2 Morphometrics used in this study..........................................................................60 3-3 Habitat of reef-associated Fraginae included in this study....................................63 3-4 Mean and 95% confidence interv als for three shell measurements.......................72 3-5 Significant differences am ong 33 species of cardiids............................................73 3-6 Principal component analysis for four morphological characters..........................75 3-7 Origin of photosymbiosis.......................................................................................76 3-8 Origin of photosymbiosis and window shell microstructure.................................79 3-9 Origin of photosymbiosis and decreased shell pigmentation................................80 3-10 Origin of photosymbiosis and maximum adult size..............................................82 3-11 Origin of photosymbiosis and habitat....................................................................83
x 3-12 Origin of photosymbiosis and microhabitat (exposure).........................................85 3-13 Origin of photosymbiosis and mean keel angle.....................................................86 3-14 Origin of photosymbiosis and m ean posterior shell flattening..............................87 3-15 Origin of photosymbiosis and mean linearized surface area to volume ratios......88
xi Abstract of Dissertation Pres ented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy EVOLUTION OF PHOTOSYMBIOSIS IN MARINE BIVALVES (CARDIIDAE: FRAGINAE) By Lisa Ann Kirkendale December 2005 Chair: Gustav Paulay Major Department: Zoology In the marine clam family Cardiidae, two closely related subfamilies (Tridacninae and Fraginae) exhibit divergent morphological responses to a similar selective pressurelight capture for photosymbiosis. All giant clams maximize light capture via a similar suite of morphological characters, in cluding mantle hypertrophy, pronounced valve gaping, and expansion of the posterior area. In contrast, the more diverse but poorly characterized Fraginae have trav elled a much different route. Photosymbiotic taxa have responded to the same selective pressure of maximizing light capture by assuming a wide diversity of morphologies, incl uding modifications to shel l shape, shell size, shell microstructure and behavior. Three mitoc hondrial (16S, COI, CytB) and one nuclear (28S) gene regions were assembled to unders tand the relationships among over thirty photosymbiotic and non-photosymbiotic Fragin ae taxa and key characters were then mapped onto this phylogeny.
1 CHAPTER 1 GENERAL INTRODUCTION Symbioses, loosely defined as one organi sm living in intimate association with another (Cowen 1983, Futuyma 1998), are a perv asive and enduring theme in the history of life (Endosymbiotic Theory, Margulis 1998, bu t see Gray et al. 1999). What are some of the features of these relationships? Host and symbiont are taxonomically diverse with representation in all three domains (Beck age 1998, Bourtzis and O'Neill 1998, Distel 1998, McFall-Ngai & Ruby 1998, Meeks 1998, Moran and Telang 1998). Symbioses span the range from loosely to more tightly coupled affiliations. In these latter tightly coupled symbioses, energy sources are directly translocated between an animal host and protist symbiont, and thus represent a trophi c level interaction (trophosymbiosis). The recent discovery of symbiont chloroplast genes in the host opisthobranch genome exemplifies the intimacy of these associations (Mujer et al. 1996). Symbioses are evolutionarily significant events as they impart host and symbiont with previously unrealized a nd therefore novel capa bilities. Evolutionary novelties, or key innovations (Mayr 1963: p.602), are im plicated in most of the adaptive diversification events known (f or a review, see Givnish 1997) . They involve a synergistic interaction between a synapomorphy (novelt y or innovation), and ecological opportunity, the result of which is the diversification or radiation of a clade (Richardson 2001). This, in turn, often allows for the occupation and exploitation of a wide range of environments, including marginal ones (C orliss et al. 1979, Reid 1990).
2 As such, they may engender taxa with cap abilities that allow adaptation to dynamic environmental conditions, especially im portant considering impending global and regional climate warming (e.g., "adaptive" bleaching hypothesis of Buddemeier and Fautin 1993). These aspects of symbiose s make them important to identify and understand. One of the best-characterized symbiotic relationships, documented among diverse host taxa in equally diverse habitats, is chem osymbiosis. In these a ssociations, the host houses chemosynthetic bacteria, which are able to exploit the chemical bond energy contained in sulfide or othe r reductive (e.g., methane) sources. The chemical bond energy is released in biologically useful forms, such as ATP and NADPH, via a series of complex enzymatic reactions. These energy sour ces are used to drive reductive steps in carbon dioxide fixation and nitrate reduction. Carbohydrat es and amino acids are produced and, in part, eventually translocated to the hosts' tissues (F elbeck et al. 1983). Hydrothermal vents host a diverse chemosym biotic fauna (Corliss et al. 1979). Host chemosymbiotic taxa are distributed across at least three invertebrate phyla, including bivalve and gastropod molluscs, vestimen tiferan, pogonophoran as well as nematode worms (Distel 1998). This chemosymbiotic fa una breaks down the sulfide energy source flowing from the vents via oxidative reactions, permitted by the high oxygen concentrations of the cold, ambient seawater (Felbeck et al. 1983). The hot vent biota is believed to be extremely ancient and prokaryot es in this community may represent early life forms (Corliss 1979, Distel 1998). Chemosymbiotic hot vent taxa provide new and exciting evidence of the importance of the symbiotic life strategy. Ho wever, a more pervasive animal-protist
3 trophic interaction has long been known in a diverse assemblage of tropical photosymbiotic taxa (Yonge 1936). Although a common theme in reef communities, nowhere are these affiliations more diverse th an in the tropical Indowest Pacific (IWP), the largest marine biogeogr aphic region on Earth (Paula y 1997). Algalor photosymbiosis is an association between a hetero trophic host, and one or more unicellular algal taxa or autotrophs (Cowen 1983). A great diversity of host taxa are known, including representatives from Foraminife ra, Porifera, Platyhelminthes, Cnidaria, Mollusca and Chordata (Zann 1980, Cowen 1983, Mujer 1996). A common theme among shallow water scleractinian corals is the obligate association with zooxanthellae or dinophytes (Douglas 1994, Veron 2000, Richards on 2001); more than 60 coral genera harbor symbionts (Cowen 1983). Most of th e research on photosymbioses have focussed on the host; however a great diversity of algal symbionts have been documented, including cyanobacteria, ch lorophytes, diatoms, rhodophytes, chrysophytes and even "free" chloroplasts (Zann 1980, Rutzler 1990, Lee and Anderson 1991). Recent work on scleractinian corals has reve aled a huge genetic diversity within the most common and abundant dinoflagellate symbiont, Symbiodinium (Rowan and Powers 1992, Rowan 1998). Spatially and temporally vari able, these host-symbiont associations yield "optimal" combinations for myriad e nvironmental and ecological conditions, thus demonstrating the adaptive pot ential of these associations (Rowan et al.1997). What are the benefits to host and symbiont? Photosym bionts produce photosyntha tes, specifically glycerol and alanine, some of which are tr anslocated to the host (Muscatine 1967, Goreau et al. 1973). These are thought to be critical for host nutri tion in the highly oligotrophic tropical environments (Cowen 1983). Simila rly, the symbionts are thought to provide
4 foraminiferan hosts with a means of recy cling metabolic waste products (Richardson 2001). The symbionts are believed to gain access to a steady supply of normally limiting nutrients (N, P and Fe) due to tightly coupl ed nutrient cycling w ith the host (Cowen 1983, Douglas 1995, Rowan 1998, Richardson 2001). For example, Yonge (1936) found that there was little leakage of phosphorus in giant cl am-zooxanthellae associations, but that large amounts of metabolic phosphorus were released fro m large tridacnines without symbionts. Most photosymbiotic taxa, including corals and shell-less opisthobranch molluscs, have exposed soft tissues and large surface area s that facilitate exposure of symbionts to light (Cowen 1983). However, other taxa such as bivalve molluscs that have opaque, hard shells covering much of their body are clearl y not Â“preadaptedÂ” for photosymbiosis, yet many are photosymbiotic. Given the obvious cons traints, how do these taxa facilitate light penetration to photosymb ionts? All but five known exam ples of photosymbiosis in bivalves occur in tropical representatives of the cockle family Cardiidae. The freshwater bivalves Anodonta cygnea and Unio pictorum harbor zoochlorellae photosymbionts, which are reported to reside in tissues that receive maximum illumination (Jones and Jacobs 1992). Only two non-cardiid photos ymbiotic marine bivalves are known, Fluviolanatus subtorta , a bizarre putatively trapezid clam from shallow, tropical Australian seas (Morton 1982) and Placoplecten magellanicus , a temperate scallop reported to harbor zoochlorellae as mantle endosymbionts. Within the cardiids, only one photosymbiotic taxon is known from temperate regions, Clinocardium nuttalli , a relatively large cockle with zoochlorellae symbionts (Jones and Jacobs 1992). All of these associations are purpor tedly facultative and not likely trophic, unlike those known
5 for tropical cardiids. However, as yet, th ese affiliations rema in relatively poorly characterized in comparison to symbioses with zooxanthellae (Jones and Jacobs 1992). The giant clams or tridacnines ( Tridacna and Hippopus ) are the best-known subfamily of photosymbiotic cardiids and ar e found throughout the shallow tropical IWP (Yonge 1936, 1981, Rosewater 1965, Schneider a nd " Foighil 1999). All tridacnines possess a similar suite of morphological and behavioral features attributed to the symbiotic lifestyle (Rosewater 1965, Yonge 1936, 1981). Zooxanthellae reside primarily in the hypertrophied posterior mantle region. To allow for the expansion of the posterior region, the anterior region of the animal ha s been concomitantly reduced. Traditionally, it was thought that zooxanthellae resided in th e hemal spaces of the hypertrophied mantle (Yonge 1936). However, it was later suspected (Mansour 1946), and has since been confirmed (Norton et al. 1992) , that zooxanthellae are housed intercellularly in an elaborate network of primary, secondary and tertiary zooxanthella r tubes. Although this elaborate duct system has not yet been desc ribed for any other photosymbiotic bivalves besides Corculum ( Janssen 1991), it is similar to those described for other molluscan taxa (opisthobranchs), and is intimately associat ed with the digestive system of the host (Trench 1979, Norton et al. 1992). Zooxanthellar transmission is horizontal in giant clams; larvae acquire zooxanthellae anew fr om the ambient environment at about 2-9 days after metamorphosis (Fitt and Tren ch 1981). Veligers without photosymbionts do not survive, suggesting an obligate photosymbi otic association and rigorous selection for the acquisition of photosymbionts (Cowen 1983). Posterior valve gaping behavior, where th e shells are widely opened to allow maximum mantle exposure, and an epibenthic existence enhance, irradiance levels to
6 photosymbionts. Enhanced exposure of soft tissues may allow for increased predation pressure, relative to the more typical narro wly cracked gape and reduced soft tissue exposure typical of most other bivalves. To comb at this threat, it is widely believed that Tridacna has evolved special defensive structures that include a sensitive optical sensory system, very rapid shell closure (via the la rge anterior adductor muscle) and directed water jetting from the siphon. Morton (1978) has shown that the defensive system of tridacnines is in operation during the day when valves are widely opened to expose photosymbionts to maximal irradiance (Cow en 1983). A fully functional gut, coupled with the added nutrition provided by the symbiont s, has resulted in the large size of all, and the truly massive size (~1 m across) of some species (e.g., Tridacna gigas ) (Schneider 1998). In stark contrast to other well-known tr ophosymbiotic bivalve lineages, are the photosymbiotic heart cockles in the subfamily Fraginae. Whereas other photosymbiotic taxa exhibit similar morphologies, such as the morphologically rather uniform Tridacninae and chemosymbiotic taxa, which exhibit convergences in many anatomical features even though they are from phyloge netically diverse stocks, photosymbiotic fragines exhibit a striking diversity of hypothesized morphological adaptations to enhance light exposure to symbionts (K awaguti 1950, Watson and Signor 1986, Ohno et al. 1995, Carter and Schneider 1997, Persselin 1998, Schneider and Carter 2001). What are some of the morphological features exhi bited by photosymbiotic Fraginae taxa? All Corculum species exhibit a similar suite of apomor phic characteristics that appear to be photosymbiotic adaptations (Gould a nd Lewontin 1978). A feature of Corculum unique among the Fraginae is a stri ctly epifaunal existence. Corculum lies posterior-side up and
7 is often extremely cryptic, as it is simila r in color to the surr ounding dead coral rubble environments it inhabits. The shell is thin, es pecially in the expose d, posterior region and is greatly flattened ante ro-posteriorly (Kawaguti 1950, Gould and Lewontin 1978, Watson and Signor 1986). This shape greatly enhances the surface area of the shell exposed to light. Another modification shared by all Corculum , but unreported thus far for all but three other fragines, is the presence of we ll-developed "window" shell microstructure (Watson and Signor 1986, Seilacher 1990, Ca rter and Schneider 1997, Persselin 1998, Schneider and Carter 2001). Corculum shell microstructure grossly resembles other fragines, but differs most significantly in ha ving: 1) less pigment, 2) microstructural windows formed on the posterior slope by the incursion of the fibrous prismatic (FPL) and dissected crossed prismatic (DCP) into the middle and inner branching crossed lamellar (BCL) and inner complex crossed lamellar (CCL) shell layers and 3) a planoconvex shape of the inner surface of posterior windows. Seilacher (1990) proposed that the transparent shell microstructure func tioned as a fiber optic system to disperse light to the zooxanthellae situ ated within the mantle of Corculum . However, reanalyses of microstructure by Carter and Schneider (1997, see also Schneider and Carter 2001) revealed that the shell microstructural adaptations more likely function as light condensing lenses. This finding is consis tent with zooxanthe llae placement in Corculum . Instead of being housed in the mantle tissue th at lies flush against th e inner surface of the exposed posterior part of the shell, as in other fragines, zooxanthe llae are most abundant in the inner surface of the extr emely flattened anterior surface of the valves (Watson and Signor 1986, Ohno et al. 1995, Carter and Sc hneider 1997). Although uncommon in other
8 fragines, light condensing shell micros tructure was documented in a putative photosymbiotic fossil cardiid, Protocardia ( Pachycardium ) stantoni (Schneider and Carter 2001). Members of the genus Lunulicardia are less-studied relatives of Corculum, but studied taxa possess a differe nt array of morphological ad aptations for photosymbiosis compared to other fragines. Well-developed ke els create a flattened posterior and provide light capture surfaces ("sola r panels") that facilitate maximal light exposure to photosymbionts through the shell (G. Paulay, pers. comm.). Additionally, some taxa have deeply involuted lunules, creating a pronoun ced concavity near the umbo; this is suspected to increase poster ior valve gaping ability by shifting the hinge axis from the anteroventral to posterodorsal (G.P. pers . comm.). Increased posterior valve gaping ability is hypothesized, in turn, to allow incr eased exposure of the mantle and thus of the symbionts. Well-developed window formation, similar to Corculum, was reported in an undescribed "juvenile" species of Lunulicardia ( L . sp. 1) but was not found in studies of adult L. retusa by Persselin (1998) or Schneider and Carter (2001). Window shell microstructure in the "juvenile" L. sp.1 of Persselin (1998) incl uded: 1) fibrous prismatic (FP) and dissected crossed prismatic (DCP) sh ell microstructure, 2) a reduction/absence of pigmentation, 3) slight convexi ties on the inner window surface. Most of the little-studied Frag inae taxa occur in the genus Fragum , clearly the most species-rich group, exhibiti ng the widest range of puta tive adaptive morphologies. Taxa possess singly or in concert, window shell microstructure, well-developed keels, posterior valve gaping behavior, some de gree of soft-body exposure and allometric change in shell shape that increases the relative area available for light capture. Fragum
9 nivale and F. mundum have fully developed windows and well-developed keels, as reported in "juveniles" of Lunulicardia sp.1 (Persselin 1998). Fragum fragum combines enhanced posterior valve gaping and well-de veloped keels with intermediate window formation (Persselin 1998). Fragum unedo and Fragum loochoanum exhibit posterior valve gaping behavior and all three of these taxa exhibit some degree of mantle splaying behavior. The photosymbiont-packed mantle is exposed along the posterior valve edges, similar to behavior exhibited by tridacnines, thus facilitating light exposure (Ohno et al. 1995). Perhaps more notable than mantle splaying behavior in F. fragum and F. loochoanum is the dense packing of zooxanthella e in the siphonal tentacles, also observed in F. mundum . Finally, the enigmatic Fragum erugatum , endemic to the hypersaline waters of Shark's Bay, Western Australia, has recently b een the subject of an extensive functional morphological study by Morton (2000). This photo symbiotic species reaches maturity in approximately 1 year and is the dominan t infaunal component of Shark's Bay, but photosymbiotic morphological attr ibutes are little known. Like F. erugatum , putative photosymbiotic adaptations are poorly understood in F. carinatum and F. scruposum, however none of these taxa have obvious window microstructure. The main focus of this dissertation resear ch is to produce a comprehensive, fully resolved phylogeny of the relationships am ong extant Fraginae taxa. A species-level phylogeny based on sequence data from multiple gene regions, including nuclear and mitochondrial markers, forms the framework to : 1) pinpoint the origin of photosymbiosis and 2) track morphological e volution related to a photosym biotic lifestyle in the Fraginae.
10 CHAPTER 2 MOLECULAR PHYLOGENETICS Introduction In the marine clam family Cardiidae, two closely related subfamilies exhibit different morphological respons es to a similar se lective pressure: light capture for photosymbiosis. While one group (the Tridacnin ae or giant clams) has responded with a similar suite of morphological adaptations, the other (the Fr aginae or heart cockles) has responded in strikingly divergent ways. All tridacnines maximize light capture via an analogous suite of morphologies and be haviors, including pronounced mantle hypertrophy and valve gaping, an epifaunal exis tence and expansion of the posterior area and scutes (shell ex tensions that Â“propÂ” up the mantle)(Yonge 1936, 1981) (see Fig. 2-1 for shell adaptations). In contrast, the mo re diverse but poorly characterized fragines have traveled a much different route, w ith photosymbiotic taxa exhibiting a wide diversity of morphological st rategies for light captur e (Bartsch 1947, Kawaguti 1950, 1968, 1983, Seilacher 1990, Watson and Signor 1986, Ohno et al. 1995). Fragines display multiple morphological solutions to the proble m of light capture with modifications to shell shape, shell size, sh ell transparency, mantle hype rtrophy and tissue exposure (Kawaguti 1950, Watson and Signor 1986, Ohno et al. 1995, Carter and Schneider 1997, Persselin 1998, Morton 2000, Schneider and Carter 2001) (see Fig. 2-1 for shell adaptations).
11 Figure 2-1 Schematic diagram of some shell adaptations in photosymbiotic Trid acninae and Fraginae. A. Tridacna squamosa , B. Fragum unedo , C. Fragum loochoanum , D. Fragum fragum , E. Lunulicardia hemicardia and F. Corculum cardissa . Dark circles denote the ligament, dashed line indicates sediment surface, arrows indicate inci dent light and hatched circles represent windows.
12 Although several morphological adaptations in this group appear to have developed in parallel, others are widely di vergent, providing evidence for the importance of historical contingency in the evolution of form (Gould and Lewontin 1978). Reconstructing the evoluti on of photosymbiosis in th e Fraginae involves mapping putative adaptations, which in turn requires a phylogenetic framework. However, current phylogenetic appraisals that in clude fragines have largely been constructed to answer higher-level systematic questions of bivalves. As such, sampling strategies have targeted wide coverage of many groups with more lim ited sampling effort within lineages (e.g., Giribet and Distel (2003) fo r the bivalvia; Keen (1980) and Schneider (1998) for cardiids). This approach, referred to as th e exemplar approach (Wheeler et al. 1993), contrasts with a more comprehensive approach where many extant taxa of a higher-level grouping (e.g., all or most species in a genus) are sampled (Meyer 2003, 2004). Although both approaches are us eful, the exemplar approach is of limited utility in the Fraginae for at least three reasons. First, exemplars are often chosen based on availability and because of this limitation, type species, although the most appropriate representatives, are often not obtained. A rela ted but somewhat different issue arises in unrevised groups, where many little-known member s lead to underestim ates of variation and overly Â“narrowÂ” taxonomic boundaries. This, in turn, influences exemplar choice and subsequent phylogenetic analys es. Second, even if all species were known and readily available, high levels of morphological vari ation can make exemplar choice difficult. Which member is best representative of th e speciose and morphologically diverse fragine genus, Fragum ? Third, single exemplars do no t test for generic monophyly.
13 Phylogenetic concerns aside, missing species (and missing character states) may lead to an incorrect understanding of patterns in studies of morphological evolution. In order to pinpoint the orig in and track the developmen t of photosymbiosis in the Fraginae, construction of a new phylogeny is required due to insufficient taxon sampling of the group in existing phylogenetic apprai sals. This chapter provides the most comprehensive molecular phylogenetic hypothesis of the subfamily to date. Nuclear and mtDNA sequence data from four gene regions were sampled from multiple representatives of all recognized genera and subgenera, as well as cardiid outgroups from numerous subfamilies. The focus of this study was to briefly review previous systematic treatments of cardiids and fr agines to facilitate compar isons with the new phylogenetic findings presented here. The evolution of photosymbiosis is treated separately (Kirkendale, Chapter 2). Higher-Level Systematics of the Cardioidea and Cardiidae The marine bivalve superfamily Cardio idea is composed of a single well-known family, the Cardiidae, known from the Late Triassic to Recent (C oan et al. 2000, Morton 2000, Schneider and Carter 2001). The appr oximately 200 members are distributed worldwide in nearshore to deep-sea envi ronments, although the bulk of species are known from shallow tropical habitats (Poutiers 1992, Morton 2000, Vidal 1994, 1997a, b, 1999, Hylleberg 2004). Traditionally comprising six subfamilies, card iid membership and subfamilial relationships have varied cons iderably among authors (e.g., the six members have not always been consis tent; compare Figs. 2-2 A&B) . Schneider (1995) recently reorganized higher-level cardiid diversity into three clades with tested representatives distributed among eleven subfamilies (Fig. 2-2C).
14 Figure 2-2 Recent cardiid (A-C) and fragin e (D) phylogenies. Dashed lines indicate extinct taxa. Clades 1-3 in C refer to major cardiid clades outlined by Schneider (1995). The re peated occurrence of Fragum in D refers to the equivocal position of this taxon. The major difference between Schneider (1998) and Persselin (1998) is the position of Fragum . Persselin places the genus as sister to Lunulicardia and Corculum ( Fragum * in D), not in a trichotomy. Both scenarios are presented by Schneider (1998).
15 Fraginae Monophyly The subfamily Fraginae was first delin eated by Stewart (1930 ) and originally included five genera and two subgenera, grouped together be cause they shared two shell characters: 1) a marked umbonal ridge and 2) subequal cardinals (T able 2-1). Although several studies have been published since the n, these largely followed StewartÂ’s original circumscription (e.g., Kafanov and Popov 1977, Keen 1980). In contrast, the most recent and comprehensive phylogenetic treatment of the subfamily Fraginae recognized eleven extant genera (and one extinct genus, Goniocardia ) united by five morphological synapomorphies: 1) small labial palps (less than one-fifth the length of animal), 2) lack of a perisiphonal suture, 3) pres ence of a functional byssus in adults, 4) poorly organized fibrous prismatic shell microstructure, discon tinuous with underlying shell layers and 5) distinct right posterior cardi nal tooth shape (Table 2-1, Fi g. 2-2 D) (Schneider 1998). Small European cardiids from the genera Parvicardium and Papillicardium were recovered as Fraginae genera, in contrast with the bulk of previous work that placed these genera elsewhere (e.g., in the Cardiinae or Ce rastodermatiinae; see Table 2-1). Consistent with earlier circumscriptions of the Fraginae but in contrast with Schneider (1998), a recent, large-scale molecular analys is of the Bivalvia did not support Parvicardium exiguum as closely related to other sampled fragines ( Corculum + Fragum ) (Giribet and Distel 2003). The Fraginae have not been the subject of a comprehensive systematic revision. This makes understanding taxonomic boundaries and controversies in the subfamily difficult. To familiarize the reader with cu rrent systematic problems, a review of
16 Table 2-1 Membership in the subfamily Frag inae. Alternate placements of genera are stated where appropriate. Genera Stewart, 1930 Keen (1980) Voskuil & Onverwagt (1989) Schneider (1998) Vidal (2001) Papillicardium Cardiinae?4 CardiinaeCerastodermatiinae X X Cerastobyssum1 X Parvicardium Cardiinae?4 CardiinaeCerastodermatiinae X X3 Trigoniocardia X X X X X Apiocardia X X X X X Lunulicardia X X X X X Corculum X X X X X Fragum X X X X X Microfragum X X X X Ctenocardia X X X X X Americardia X X X X X Afrocardium X X CardiinaeX Goniocardia2 X 1 Cerastobyssum is considered a subgenus of Parvicardium by most authors (e.g., van Aartsen and Goud 2000). 2 Goniocardia is the only entirely extinct taxon. 3 Papillicardium is considered a subgenus of Parvicardium 4 The question marks reflect Stew art's uncertainty with placing Parvicardium in the Cardiinae. presently recognized genera and species, broken down into th ree major groups, is outlined below. Parvicardium and Papillicardium : Sister to all other Fragines? Conflicting views exist regarding the earli est diverging Fragin ae genera. Although Parvicardium and Papillicardium are supported as sister to a ll fragines by some authors, these same lineages are not considered frag ines and are placed in entirely different cardiid subfamilies by others (e.g., compar e Keen 1980 with Voskuil and Onverwagt 1991 and Stewart 1930 with Schneider 1998, Tabl e 2-1). Moreover, membership in, and relationships among, the earliest diverging fragine genera are either controversial or ill defined. For example, sampling of just two species by Schneid er (1998) yielded a
17 paraphyletic Parvicardium , with P. siculum sister to Cerastobyssum hauniense and P. exiguum sister to all remaining fragines. Although Papillicardium is generally consider ed monotypic (the sole representative being P. papillosum ), a number of species are recognized in the genus Parvicardium sensu stricto (this notation will be us ed throughout the remainder of this paper to refer to species in the subgenus Parvicardium only, thus excluding Papillicardium )(Table 2-2). All but two of these ( Parvicardium turtoni from South Africa and P. pinnulatum from the temperate west Atlantic) are distributed in nearshore temperate regions of Europe, with the grea test diversity found in the Mediterranean. Based on analysis of internal and external shell characters, Voskuil and Onverwagt (1989) and van Aartsen a nd Goud (2000) considered Parvicardium sensu lato ( Parvicardium + Papillicardium ) to be composed of eight species. The latte r study is the most recent and thorough treatment of small European cardiids, with comparisons among conspecifics facilitated by SEM analysis of prodissoconchs.
18 Table 2-2 Species and distribut ion of recognized Fraginae with asterisks denoting types. Taxa Authority/year Distribution Fragum fragum * (LinnÃ©, 1758) IWP Fragum scruposum (Deshayes, 1855) IWP Fragum loochoanum Kira, 1959 IWP Fragum carinatum (Lynge, 1909) IWP Fragum mundum (Reeve, 1845) IWP Fragum nivale (Reeve, 1845) Indian ocean Fragum unedo (LinnÃ©, 1758) IWP Fragum erugatum (Tate, 1889) Western Australia Fragum sueziense (Issel, 1869) IWP Lunulicardia retusa (LinnÃ©, 1767) IWP Lunulicardia hemicardia (LinnÃ©, 1758) IWP Lunulicardia aequale (Deshayes, 1855) IWP Corculum cardissa (LinnÃ©, 1758) IWP Corculum dionaeum (Broderip & Sowerby, 1829) IWP Corculum monstrosum (Chemnitz 1782) IWP Corculum obesum (Bartsch, 1947) IWP Corculum humanum (Chemnitz, 1782) IWP Corculum laevigatum (Bartsch, 1947) IWP Corculum aselae (Bartsch, 1947) IWP Trigoniocardia granifera* (Broderip & Sowerby, 1829) East Pacific Trigoniocardia antillarum (dÂ’Orbigny in Ramon de la Sagra, 1846) Caribbean Trigoniocardia ceramidum (Orbigny, 1842) Caribbean Apiocardia obovale* (Sowerby, 1833) East Pacific Americardia biangulata (Sowerby, 1829) East Pacific Americardia media* (LinnÃ©, 1758) Caribbean1 Americardia speciosa (Adams & Reeve, 1850) St. Helena Island Americardia planicostata (Hertlein & Strong,1947) East Pacific Ctenocardia fornicata (Sowerby, 1841) IWP Ctenocardia victor (Angas, 1872) IWP Ctenocardia translatum (Prashad, 1932) IWP Ctenocardia perornata Iredale, 1929 IWP Ctenocardia symbolica* Iredale, 1929 IWP Ctenocardia fijianum Vidal & Kirkendale, in press IWP Ctenocardia gustavi Vidal & Kirkendale, in press IWP Microfragum subfestivum Vidal & Kirkendale, in press IWP Microfragum festivum (Deshayes, 1855) IWP Parvicardium minimimum (Philippi, 1836) Mediterranean Parvicardium exiguum* (Gmelin, 1791) Mediterranean Parvicardium scriptum (Bucquoy, Dautzenbery & Dollfus, 1892) Mediterranean Parvicardium vroomi Van Aartsen, Menkhorst & Gittenberger, 1984 Mediterranean Parvicardium trapezium Cecalupo & Quadri, 1996 Mediterranean Papillicardium papillosum (Poli, 1795) Mediterranean Parvicardium scabrum (Philippi, 1844) Mediterranean Parvicardium pinnulatum (Conrad, 1831) East Atlantic Parvicardium turtoni (Sowerby, 1894) South Africa 1 type locality as Habana, Cuba selected by McLean 1839 pg. 167 and not Indian ocean, as stated originally by LinnÃ©.
19 The name Parvicardium has also been applied to some Indo-west Pacific (IWP) species (e.g., Oliver 1992 for sueziense ), but it is now widely accepted that tropical IWP representatives belong in the genus Fragum . Inclusion of P. turtoni from South Africa brings the total number of species in the genus to nine (Table 2-2). The Â“ Trigoniocardia Â” and Â“ Ctenocardia Â” groups Two other subclades recognized in the Fraginae are: 1) the Â“ Trigoniocardia Â” group that includes species in the genera Trigoniocardia sensu stricto, Apiocardia (recognized as a subgenus of Trigoniocardia most recently by Keen (1980) and Vokes (1989)) and the fossil taxon Goniocardia, and 2) the Â“ Ctenocardia Â” group composed of the genera Ctenocardia sensu stricto, Americardia and Microfragum (the latter two are recognized as subgenera of Ctenocardia by Voskuil and Onverwagt 1989 and Keen 1980, respectively) (Schneider 1998, Persselin 1998). The Â“ Trigoniocardia Â” group, united by four synapomo rphies, was recovered as a well-supported sister clade to the remaining fragines ( Lunulicardia + Corculum, Fragum and the Â“ Ctenocardia Â” group) (Fig. 2-2D) by Schneider (1998). Trigoniocardia sensu lato includes four recognized species with distributions in the East Pacific and western Atlantic/Caribbean (Table 2-2). The Â“ Ctenocardia Â” group, supported by three morphological synapomorphies, is the most diverse genus of fragines with three recognized subgenera and thirt een species (Table 2-2). The bulk of representatives are restricted to the IWP except for one subgenus, Americardia, known from the Americas. Ctenocardia sensu lato was recovered in a trichotomy with Fragum and Lunulicardia + Corculum (Fig. 2-2D) by Schneider (1998). Member s of this genus have a wide depth range, with many species found in much greate r depths than known for any other fragine
20 (e.g., both C. victor and C. fornicata are often found on steep fore-re ef slopes at depths of 100 m or more). In contrast with Schneider (1998) and Voskuil and O nverwagt (1989), the bulk of earlier taxonomic work (Stewart 1930, Clench and Smith 1944, Keen 1951, 1980, Olsson 1961 and Popov 1977) recognized Americardia as a subgenus of Trigoniocardia . These appraisals do not unite mo rphologically similar groups from disjunct locales (e.g., Ctenocardia from the IWP and Americardia from the Americas) but instead group geographically proximate but highly divergent forms (e.g., Apiocardia and Trigoniocardia with Americardia ). Known Photosymbiotic Representatives: Fragum, Lunulicardia and Corculum Reef-associated spec ies in the genera Fragum , Lunulicardia and Corculum include the most morphologically divergent cardiid s, with many confirmed photosymbiotic members (Kawaguti 1950, 1968, 1983, Trench et al. 1981, Ohno et al. 1995). However, these fragines are ge nerally poorly known ( Corculum being the obvious exception), often with a sparse fossil record due to their occurre nce in remote, little-studied settings in the tropical Pacific and Indian oceans. Maruyama et al. (1998) and Giribet and Distel (2003) both included three IWP fragines, Fragum unedo , F. fragum and Corculum cardissa, in their molecular phylogenetic analyses and recovered similar topologies, with C. cardissa sister to F. unedo + F. Fragum . Persselin (1998) proposed that a paraphyletic Fragum likely gave rise to Lunulicardia and Corculum and placed these three genera in a subclade within the Fraginae. Schneider (1998) outlined a scenario that grouped Lunulicardia + Corculum based on five synapomorphies and excluded Fragum (Fig. 22D).
21 Although revisionary work in the fragines is generally sparse, one genus has received considerable attention. Corculum is the most widely studi ed fragine, both from a morphological and taxonomic standpoint, and has long been recognized as the "other" photosymbiotic clam (Bartsch 1947, Ka waguti 1950, Gould and Lewontin 1978, Watson and Signor 1986, Seilacher 1990, Ohno et al. 1995, Carter and Schneider 1997, Persselin 1998, Farmer et al. 2001, Schneider and Carter 2001). Seven species are recognized and these are widely distributed throughout the IWP in shallow, coral rubble habitats (Bartsch, 1947) (Table 2-2). These species are differentiated on the basis of external shell characters: 1) whether the keel margin is sm ooth or denticulate, 2) whether the posterior or anterior side is concav e or convex, 3) presence of pronounced ribbing and 4) size. Kawaguti (1950) challenged this division and suggested that Corculum represents one highly polymorphic species split into groups that follow an ontogenetic sequence. Although the sympatry of many Corculum species in the Philippines supports this latter contention, the diminutive C. dionaeum as well as a western Indian ocean form (mentioned briefly by Bartsch 1947: 223 and placed provisionally with C. cardissa ), are geographically separated from the bulk of other Corculum . Due to the poor opportunities for connectivity given a short-lived larval st age (2 days in tested individuals; see Kawaguti 1950) these insular populations ma y be distinct (e.g., reciprocally monophyletic lineages or ESUs, Moritz 1994), sim ilar to tested insular Pacific gastropods with limited larval dispersal capabilities (Kirkendale and Me yer 2004, Meyer et al. 2005). The most species-rich and poorly understood group, exhibiting the widest range of putative adaptive morphologies for a photosymbiotic lifestyle, is Fragum (Ohno et al. 1995, Persselin 1998). Although the genus has never been revised, a handful of
22 representatives have been included in mol ecular phylogenetic treatments (Giribet and Distel 2003, Maruyama et al. 1998) or have b een the focus of microstructural analyses (Persselin 1998, Schneider and Carter 2001). Ni ne species are recognized (Table 2-2, following Hylleberg 1994, who in turn base d it on Moore et al. (1969), Wilson and Stevenson (1977), Koyama et al. (1981), O liver (1992) and Lamprell and Whitehead (1992)) with clear morphological differe nces among most recognized members. Boundaries among F. loochoanum, F. scruposum and F. carinatum , hereafter referred to as the Â“25-rib groupÂ”, are mo re confusing (J. Vidal and G. Paulay pers.comm.). This assemblage shares a number of characters that differentiate them from other IWP fragines, such as the presence of approximately 25 ribs, small adult size, consistent shell shape and relatively deep in terribs. However, the presence (e.g., in the Museum National dÂ’Histoire Naturelle, Paris) of a wide range of intermediates, that likely constitute a number of un recognized species, makes sort ing insular populations into consistent taxonomic groups difficult. The proper position of two species, F. sueziense and F. erugatum, has also been controversial. Although both are now gene rally accepted as members of the genus Fragum , F. erugatum has been placed in five differe nt genera (see Hylleberg 2004:502), while F. sueziense has been allied with 6 different ge nera or subgenera (see Hylleberg 2004:793). Fragum representatives inhabit a diversity of shallow tropical microhabitats from high-energy, exposed outer reef flats to lower-energy, more protected inner reef flats to muddy sands in large, shallow, ma ngrove-lined bays to deep, often turbid lagoonal settings and even shallow hy persaline coral sand environments.
23 Lunulicardia , often considered a subgenus of Fragum (Wilson and Stevenson 1977), is a small but little-studied assemblage that has never been formally revised although members have been included in recent microstructural anal yses (Schneider and Carter 2001, Persselin 1998). Wilson and St evenson (1977) recognized two species in Australia: L. hemicardia and L. retusa (Table 2-2). Although symp atric and distributed in shallow, clean, coral sand environments throughout the IWP, L. hemicardia is more often recovered inshore of, and in slig htly shallower habitats than, L. retusa . A number of specific systematic questions regarding the Fragin ae are addressed in this study. Briefly, are the Fraginae monophylet ic? If not, what members are included and what species/groups are excluded? Are Parvicardium and Papillicardium sister to the remainder of the fragines? How are these two gene ra related and what is their membership? Generic monophyly is largely untes ted at this time; do recognized genera form well-supported clades? What are the re lationships among genera, and how do they compare to previous taxonomic appraisals? Are sampled members of the American genus Americardia more closely related to mor phologically similar species of Ctenocardia from the IWP or to other American fragines (e.g., Trigoniocardia and Apiocardia )? Is Fragum recovered in a polytomy with (S chneider 1998) or sister to, Lunulicardia and Corculum (Schneider 1998, Persselin 1998)? Comprehens ive sampling of species and populations in little-known Fragum will facilitate proper placement of this diverse and difficult genus. Are the enigmatic F. sueziense and F. erugatum supported as members of Fragum or more closely related to non-fragine cardiid s? Finally, little studied and suspected photosymbiotic species in the divers e IWP (e.g., members of the genera Fragum and Lunulicardia ) were targeted for geographic/populat ion-level sampling to test species
24 boundaries. Morphologically divergent Corculum from a wide geographic range are included to test whether Corculum is comprised of numerous species (Bartsch 1947), or better described as a single, ecophenot ypically variable one (Kawaguti 1950). Materials and Methods Specimen Acquisition Fraginae were collected worldwide, resulti ng in genetic material of almost 60% of ingroup species with representatives sampled from all extant gene ra and subgenera (as specified by Schneider 1998)(Tables 2-2 and 2-3). Outgroups included representatives from four cardiid subfamilies: Laevicar diinae, Protocardiinae, Cardiinae and Lymnocardiinae (Table 2-3). Two individuals were seque nced per species, where possible, for each of four gene regions: three mitochondrial (Cytochrome oxidase I (COI), 16S, Cytochrome B (CytB)) and one nuc lear (28S rDNA). All samples were fixed in ethanol for molecular analyses, and all i ndividuals (shells and une xtracted tissues) are housed at the Florida Museum of Na tural History (UF) (Table 2-3). DNA Extraction and PCR Total genomic DNA was obtained from ethanol -preserved muscle tissue (foot or if the animal was <1 cm, entire body) us ing DNAzol (Chomczynski et al. 1997) methodologies at one-half suggested volumes with extended digestion times (1 day-1 week)(Molecular Research Center, Inc.). Prim ers D1F and D6R were used to amplify and sequence the D1-D3 domains of 28S rDNA fo r most species, but occasionally D2F was used in place of D1F (Park and " Foighil 2000 ; Table 2-4). Universal primers were used for COI (Folmer et al. 1994), 16S (Palumbi 1996) and CytB (Kocher et al. 1989), with specific COI primers designed to target taxa in the genera Fragum , Lunulicardia and Corculum (Table 2-4). PCR cocktails included 1 ÂµL of genomic DNA template, 5 ÂµL of
25 Table 2-3 Sampling localities and voucher information for a subset of Fraginae representatives sequenced for th e Â“coreÂ” phylogeny. Numbers following names denote the unique genomic extraction number. Taxa Locality Accession numbers INGROUP Parvicardium exiguum Mediterranean from GenBank Parvicardium papillosum278 Mediterranean UF374115 Parvicardium minimum194 Sweden UMICH265486 Parvicardiumvroomi294 Mediterranean UF374116 Parvicardium scriptum283 Mediterranean UF374117 Parvicardium sp. LaHerra 299 Mediterranean UF374118 Americardia media388 Caribbean UF347556 Americardia biangulata331 Panama UF351615 Trigoniocardia granifera334 Panama UF359687 Apiocardia obovale398 Panama UF351671 Ctenocardia victor3 Guam UF288935 Ctenocardia fornicata18 Tanzania UF286471 Ctenocardia gustavi311 Papua New Guinea UF351689 Microfragum festivum471 Northeastern Australia UF374119 Fragum sueziense56 New Caledonia UF299263 Fragum erugatum134 Western Australia UF299293 Fragum fragum 61 Rangiroa UF299282 Fragum unedo129 Western Australia UF299291 Fragum carinatum318 Papua New Guinea UF351691 Fragum loochoanum121 Guam UF299448 Fragum scruposum316 Northeastern Australia UF374114 Fragum aff . mundum377 Tuvalu UF348016 Fragum mundum78 Hawaii UF296894 Corculum cardissa9 Sulawesi UF280389 Lunulicardia retusa21 Western Australia UF291497 Lunulicardia hemicardia136 Papua New Guinea UF299269 OUTGROUP Laevicardium sp.79 Zanzibar UF285613 Papyridea semisulcata80 Florida UF286647 Fulvia australis110 Palawan UF286335 Frigidocardium sp.13 8 Florida UF294008 Acanthocardia echinata204 Sweden UMICH265485 Cerastoderma edule300 Europe UF374113 Papyridea sp.336 Panama UF351597 Microcardium sp.341 Panama UF351592 Cerastoderma glaucum lamarcki i Sweden UMICH265488
26 Table 2-4 Primers used for phylogenetic r econstructions of Fr aginae relationships. Primer name /gene region Sequence 28S-D1F 5'-GGAACTACCCCCTGAATTTAAGCAT-3' 28S-D2F 5'-TCAGTAAGCGGAGGAA-3' 28S-D6R 5'-CCAGCTATCCTGAGGGAAACTTCG-3' LCO1490 5'-GGTCAACA AATCATAAAGATATTGG-3' HCO2198 5'-TAAACTTCAGGGTGACCAAAAAATCA-3' FRAG1-LCO 5'-TCA TTT AGW ATY ATK ATY CGW AC-3' FRAG2-LCO 5'-TCT TTT AGR RTW ATA ATY CGW AC-3' FRAG1-HCO 5'-GAC CAA AA A ATC ARA ANA RAT G-3' 16Sar 5'-CGCCTG TTTATCAAAAACAT-3' 16Sbr 5'-GCCGGTCTGAACTCAGATCACGT-3' CytB 5'-AAAAAGCTTCCATCCA ACATCTCAGCATGATGAAA-3' CytB 5'-AAACTGCAGCCCCTCAGAATGATATTTGTCCTCA -3' 10X buffer, 5 ÂµL 10 mM dNTPs, 2 ÂµL of 10 ÂµM solution of each primer, 2-4 ÂµL of 25 mM MgCl 2 solution, 0.2 ÂµL TAQ, 2.5 Âµl DMSO brought up to a total volume of 50 ÂµL with ddH20. Reactions were run for 35-40 cycl es with the following parameters for the mitochondrial genes: an initia l 1-2.5 min. denaturation at 95 C; further denaturation at 94-95 C for 40 sec., annealing at 38-44 C (COI, CytB), 48-55 C (16S) for 35-45 sec. and extension at 72 C for 1-3 min. (with larger fragments requiring longer extension times). The 28S profile followed Park and " Foighil (2000) with 36 cycles: denaturation for 4 min. at 94 C followed by 40 sec. at 94 C, annealing for 40 sec. at 55 C, and extension for 1.45 at 72 C and 10 min. at 72 C. The addition of a Â“hot startÂ” step before PCR (10 min. at 99 C) was used for trouble taxa and/ or gene regions. The PCR product was electrophoresed, stained, and photodocumen ted. Multiple PCR products, indicated by double bands, were subjected to increased annealing temperatures during subsequent rounds. Successful PCR products were clean ed for cycle sequencing using Wizard Preps (Promega) following described protocol and th en visualized. Approximately ninety-five percent of sequences were generated us ing an ABI Prism 377 automated sequencer
27 following manufacturers' r ecommendations and utiliz ing ABI Big Dye with DyeDeoxyTermination protocols (P erkin Elmer). A small subset of CytB sequences were generated with a Beckman CEQ 8000 (Beckma n-Coulter) automated sequencer following manufacturerÂ’s recommendations. Alignment & Molecular Analyses Sequences were initially aligned by eye during editing in Sequencher 3.1.1 (Genes Codes). COI and CytB sequences were aligned ea sily due to an absence of indels. Default parameters in Clustal X (version 1.81) were used to aid in alignment of the 28S gene region (Thompson et al. 1997). Two areas with in hypervariable loop regions of 28S and 16S (100 bp and 150 bp and 126 bp and 58 bp, respect ively) were especially difficult to align among major clades, and topologies result ing from analyses with and without these regions were compared. Fine-scale resolution was lost among closely related species after the removal of hypervariable regions, while major lineages were not altered. Thus, variable sites in 28S and 16S were included in subsequent an alyses. In all analyses, gaps were treated as missing and character states were unordered. Partition-homogeneity tests, implemented in PAUP* 4.0b10 (Swofford 2002), were run to test for significant differences among all four gene regions. Significant differenc es were detected among all tested gene regions and, as a result, topol ogies were generated and compared for each gene region to facilitate visual examinati on of possible conflict s. The single-locus mitochondrial datasets were congruent, di ffering mainly in resolution due to rate variation, thus these th ree datasets were combined into a single mitochondrial dataset for analysis. Full datasets are av ailable from LAK upon request. Molecular analyses for phylogenetic rec onstructions were performed using a twotiered approach. First, all gene regions samp led from all available Fraginae species, as
28 well as outgroup representatives, were a ssembled to construct an overall or Â“coreÂ” phylogeny, largely to test subfamily and gene ric monophyly. These fi rst-tier analyses were performed using maximum parsim ony (MP) and maximum likelihood (ML) methods (the latter where possible) imple mented in PAUP* 4.0b10 (Swofford 2002) and Bayesian analyses (with burnin excluded after runs) implemented in MrBayes 3.1.1 (Ronquist & Huelsenbeck 2003). ML and Baye sian analyses were run at the UF Phyloinformatics High Performance Computi ng Cluster. These analyses were run on three datasets; mitochondrial, nuclear a nd combined nuclear and mitochondrial. Unordered and user-specified 3:1, 5:1 and 10:1 transversion biases were assigned in parsimony to correct for saturation. However, b ecause resultant topologies from variable weighting schemes did not differ from those generated using equally weighted datasets, equal weighting was employed in later anal yses. ModelTest 3.7 (Posada and Crandall 1998) was used to determine the appropriate model of molecular evolution for all other analyses (all GTR+I+G for both AIC and LRT). Tree robustness was assessed using bootstraps (100-1000 replicates, MP and ML) and posterior probabilities (Bayesian). A second-tier of molecular analyses was performed to test generic and species boundaries in all sampled species of the IWP genera Fragum , Lunulicardia and Corculum included in core analyses (Table 2-3). Additional sp ecimens (2-20 per species when available) were sequenced for the COI ge ne region (see author for full dataset) and neighbor-joining (GTR) and maximum parsimon y analyses were performed in PAUP* with bootstrap values generated from 1000 replicates. The neighbor-joining method was chosen because the dataset was large and th is reduced computation time. Congruence was established by comparison with the core phylogeny.
29 Results General FindingsSequence Data The bulk of ingroup fragine species, with representatives sampled from all extant genera and subgenera, were sampled for core phylogenetic reconstructions. Moreover, nine species from seven genera, representing four cardiid subfamilies, were included as outgroups for rooting purposes (Tables 2-2 and 2-3). Two individuals were sequenced per species, where possible, for each of four gene regions, resulting in a total genetic dataset of 3016 bp (1413 for 28S, 598 for 16S, 746 for COI and 400 for CytB). This final dataset contained 25 (roughly 60%) ingroup species ; 72% of ingroup re presentatives (and multiple individuals of a species) had comp lete mitochondrial datasets, while 60% had nuclear representation. Moreover, COI from fifty-six individuals of IWP Lunulicardia, Corculum and Fragum was sampled from available geographic endpoints to test for substructuring. General FindingsPhylogenetic Methods Bayesian methods provided the greatest resolution and suppor t (via posterior probabilities) compared with other met hods including maximum parsimony, neighborjoining and maximum likelihood. Although conflict ing results were not generated for MP and ML analyses of datasets analyzed using Bayesian methods, support was notably lacking at deeper nodes (contrast posterior probabilities with bootstrap support in Figs. 23 through 2-6) and resulted in poor resolution of certain groups. Due to concerns with elevated posterior probabilitie s (Suzuki et al. 2002), additio nal Bayesian runs (3-5) and longer generations (5000000 + 10000000) for core phylogenetic reconstruction were run to test whether major clades were consiste ntly recovered and we ll-supported. However, even with the increased resolution at deep phylogenetic levels provided by Bayesian
30 analyses, incongruent topologies were recovered for the three datasets (28S, mitochondrial and combined). Not surprisingl y, this occurred at poorly supported nodes (Figs. 2-3 through 2-6). For example, Clades II and III were recovere d as sisters in the mitochondrial dataset, which was, in turn, sist er to a clade of outgroup species with both Clades I and II more distantly related (Figs. 2-3 and 2-5). However, in the combined dataset a deep polytomy of four clades (C lade I, Clade II, Clade III and an outgroup clade), sister to Clade II, was best suppor ted given extremely low posterior probabilities (52, 54, Figs. 2-3 and 2-6). Poor resolution at deeper phylogenetic leve ls was a result of four main issues: 1) alignment uncer tainty, 2) incomplete taxon sampling, 3) highly divergent sequences of certai n species and 4) poor signal/mar ker choice for higher level reconstructions. Given these concerns, phyloge netic inference was restricted to wellsupported nodes in comprehensively sampled combined and mitochondrial phylogenies analyzed by Bayesian methods. The 28S data set is presented for comparison with the other two phylogenies, however poor taxon samp ling, coupled with alignment difficulties in some groups, limit the utility of this gene region for robust phylogenetic reconstruction. Higher-Level Phylogenetic Findings Four major clades of fragines were well supported in Bayesian analyses of mitochondrial and combined datasets (Figs. 2-3 through 2-6). These four main clades correspond to: 1) a group that includes the majority of te sted members from the genus Parvicardium , including P. exiguum (Clade I), 2) SchneiderÂ’s (1998) Â“ Trigoniocardia Â” and Â“ Ctenocardia Â” groups (Clade II) except C. victor , 3) all species in the genera Fragum , Corculum and Lunulicardia (Clade III) and 4) two highly divergent representatives of Parvicardium and Papillicardium (Clade IV) (Figs. 2-3 through 2-6).
31 Fraginae Monophyly The four major clades (C lades I-IV), comprising all sampled members of the subfamily Fraginae, were not recovered as monophyletic in any analyses of the core dataset (Figs. 2-3 throug h 2-6). Two species of Parvicardium sensu lato ( Papillicardium papillosum and Parvicardium minimum ) were highly divergent from all other tested ingroup and outgroup cardiids (Clade IV). Ctenocardia victor was similarly divergent and consistently fell with distantly relate d outgroup species instead of with congeners. Figure 2-3 Overview of major clades and s upport (posterior probabilities) of A. 28S + mtDNA (16S, COI, CytB) and B. mtDNA only Bayesian phylogenies.
32 91/100 96/95 100/100 72/ 93/100 100/100 99/100 93/86 97/100 98/100 100/100 94/9990/92 98/9865/99OUTGROUP CLADE II Trigoniocardia + Ctenocardia CLADE I Parvicardium CLADE III Fragum, Corculum + Lunulicardia Parvicardium sp. Parvicardium vroomi Parvicardium vroomi Parvicardium scriptum Papillocardium papillosum Papillocardium papillosum Fulvia australis Papyridea semisulcata Papyridea sp. Fragum sueziense Fragum sueziense Fragum scruposum Fragum scruposum Fragum scruposum Fragum loochoanum Fragum carinatum Fragum carinatum Fragum fragum Fragum unedo Fragum unedo Fragum mundum Fragum mundum Fragum mundum Corculum cardissa Corculum monstrosum Lunulicardia hemicardium Americardia media Americardia media Ctenocardia gustavi Nemocardium sp.100 100 100 84 100 90 86 100 100 100 100 99 100 100 100 100 100 100 67 100 100 100 88 100 95 100 OUTGROUP CLADE IV Papillicardium Figure 2-4 Bayesian 50% majority t opology for the 28S gene region. Posterior probabilities above branches, likeliho od/parsimony bootstrap values below branches.
33 100 52 88 100 100 89 98 93 56 88 60 99 99 100 100 97 93 100 96 93 100 100 100 99 100 97 100 99 100 100 60 100 98 68 58 100 88 83 100 100 100 96 100 100 56 100 100 100 74 100 87 100 100 100 100 100 100 100 99 100 100 100 100 100 100 99 63 100 100 100 100 100 82 100 99 95 100 100 83 100 58 100 100 100 100 100 99 100 100 99 78 84 93 100 100 100 100 100 100 100 100 OUTGROUP CLADE II Trigoniocardia + Ctenocardia CLADE I CLADE III Fragum, Corculum + LunulicardiaParvicardium sp. Parvicardium vroomi Parvicardium scriptum Papillocardium papillosum Fulvia australis Papyridea sp. Fragum sueziense Fragum scruposum Fragum loochoanum Fragum carinatum Fragum fragum Fragum unedo Fragum aff. mundum Fragum mundum Corculum cardissa Corculum monstrosum Lunulicardia hemicardium Americardia biangulata Americardia media Ctenocardia gustavi Nemocardium sp. Parvicardium minimum Ctenocardia fornicata Microfragum festivum Trigoniocardia granifera Apiocardia obovale Fragum erugatum Lunulicardia retusa Ctenocardia victor Laevicardium spp. Frigidocardium sp. Cerastoderma edule Cerastoderma glaucum Acanthocardia spp. CLADE IV Parvicardium sensu stricto Papillicardium + P. minimum Ctenocardia victor Parvicardium exiguum OUTGROUP Figure 2-5 Bayesian 50% majority topol ogy for the mitochondrial (16S, COI, CytB) dataset. Posterior probabilities above branches,maximum parsimony bootstrap values below branches.
34 100 100 100 100 100 54 52 100 99 98 100 100 98 99 90 100 100 100 100 100 100 86 100 100 100 100 51 99 95 94 100 99 100 100 100 100 86 100 99 100 100 100 86 100 100 93 100 100 100 100 100 100 100 100 88 96 100 100 100 100 100 100 100 100 100 100 OUTGROUP CLADE II Trigoniocardia + Ctenocardia CLADE III Fragum, Corculum + LunulicardiaParvicardium sp. Parvicardium vroomi Parvicardium scriptum Papillocardium papillosum Fulvia australis Papyridea spp. Fragum sueziense Fragum scruposum Fragum loochoanum Fragum carinatum Fragum fragum Fragum unedo Fragum aff. mundum Fragum mundum Corculum cardissa Corculum monstrosum Lunulicardia hemicardium Americardia biangulata Americardia media Ctenocardia gustavi Nemocardium sp. Parvicardium minimum Ctenocardia fornicata Microfragum festivum Trigoniocardia granifera Apiocardia obovale Fragum erugatum Lunulicardia retusa Ctenocardia victor Laevicardium sp. Frigidocardium sp. Cerastoderma edule Cerastoderma glaucum Acanthocardia spp. Parvicardium sensu stricto Ctenocardia victor Parvicardium vroomi Laevicardium sp. 56 84 86 96 100 100 78 61 50 88 94 57 93 100 99 100 98 97 100 99 99 86 100 100 100 63 60 100 94 50 55 90 99 64 99 98 90 100 100 92 99 100 100 84 100 Parvicardium exiguum CLADE I CLADE IV Papillicardium + P. minimum OUTGROUP Figure 2-6 Bayesian 50% majority topology fo r the mitochondrial (16S, COI, CytB) and 28S datasets. Posterior probabilities above branches, maximum parsimony bootstrap values below branches. Sequence quality and alignments were verified for multiple individuals of these three aberrant species, confirming that highly dive rgent sequences were not artifacts. Although the Fraginae were found to be polyphylet ic, circumscribing a new Fraginae was complicated by poor resolution at deeper node s that yield incongruent mitochondrial and combined topologies. For example, an outgr oup cardiid assemblage was sister to the fragine Clades II and III, not ot her fragines (e.g., either Clade I or IV), in analyses of the
35 mitochondrial dataset (Fig. 2-5). In contra st, Clade III was recovered as the poorly supported sister to Clades I and II in the combined dataset (Fig. 2-6). The Genus Parvicardium sensu lato: Clades I and IV Bayesian mitochondrial and combined phyl ogenies consistently recovered high support for two clades of members from Parvicardium sensu lato (Clades I and IV in Fig. 2-4 & 2-5). Clade I was com posed of four species of Parvicardium : an undescribed but divergent taxon known from one individual ( P . sp. 1), P. vroomi , P. scriptum and P. exiguum. Species boundaries between P. vroomi and P. scriptum were unclear; although P. vroomi was recovered as monophyletic based on mitochondrial analyses (Fig. 2-4), Bayesian analysis of the combined datasets suggested a paraphyletic P. vroomi (Fig. 2-5). The two Parvicardium species from Clade IV, Parvicardium minimum and Papillicardium papillosum, were extremely divergent from all ingroup and outgroup cardiids, and from each other for all gene regions tested. The Â“ CtenocardiaÂ” and Â“ TrigoniocardiaÂ” groups: Clade II Clade II was recovered as a well-supported fragine subclade in combined and mitochondrial phylogenies, composed of all Ctenocardia and Trigoniocardia group members, except for the abbarent C. victor that consistently fell with out groups (Figs. 2-4 & 2-5). In both mitochondrial and combined Bayesian analyses, repr esentatives of Americardia were more closely related to sampled members of Trigoniocardia , not included conspecifics. Generic monophyly of Ctenocardia and Trigoniocardia was not supported be cause of a polyphyletic Ctenocardia (Figs. 2-4 & 2-5). Trigoniocardia granifera, Apiocardia obovale and Americardia biangulata clustered to the exclusion of Americardia media in combined and mitochondrial Bayesian analyses. Thus, Americardia was not supported as m onophyletic. Boundaries in
36 Apiocardia and Microfragum could not be tested; only one repr esentative of each was included in the study. Fragum , Lunulicardia and Corculum : Clade III Clade III, composed of the IWP photosymbiotic genera Fragum , Lunulicardia and Corculum, is the largest subclade and consisten tly highly supported and well resolved in all analyses (Figs. 2-3 thr ough 2-6). Within this clade, three well-supported subclades were recovered in combined and mitochondria l phylogenies: 1) a subclade of earliest diverging members; F. sueziense and F. erugatum , 2) the Â“25-ribÂ” subclade and 3) a subclade uniting Fragum mundum, Corculum, Lunulicardia, F. fragum and F. unedo . Within the latter subclade, two additional subclades were resolved in the combined dataset: 1) F. Fragum and F. unedo and 2) a group including F. mundum and Corculum, which was in turn sister to Lunulicardia (Fig. 2-6). The second major Fragum complex recovered, the Â“25-ribÂ” group, includes F. loochoanum , F. scruposum , F. carinatum and F. aff. mundum . F. aff. mundum was sister to F. carinatum, which was in turn, sister to the remaining 25-rib members. Corculum and Lunulicardia were supported as monophyletic in all analyses, while Fragum was paraphyletic (Figs. 2-3 through 2-6). The m itochondrial topology differed only slightly from combined analyses with respect to relationships in Clade III. F. unedo and F. fragum were not recovered as sisters, as supported by combined analyses. Instead, F. unedo was sister to all other members of its subclade (compare Fig. 2-3 A, B). SpeciesÂ’ Boundaries in IWP Corculum , Lunulicardia and Fragum Population-level COI specific analyses recovered all Fragum species (Table 2-2) as reciprocally monophyletic lineages (Fig. 2-7) . Because of the difficulties with species recognition using morphological characters, the 25-rib group was targeted for intense
37 molecular scrutiny (n=25). A deep and wellsupported split occurr ed uniting specimens identified as F. scruposum from Papua New Guinea and Australia (n=7), to the exclusion of a group provisionally recognized as F. loochoanum from Okinawa, Papua New Guinea (PNG), Tuvalu, French Polyne sia and Micronesia (n=17) (F ig. 2-7). Population-level sampling in the western and central Pacific re vealed geographic struct uring in four other species in the genus Fragum : F. unedo, F. fragum , F. sueziense , F. mundum . To test species concepts within Corculum , COI sequence data from C. cardissa, C. aselae and C. monstrosum were analyzed. Individuals identified as C. aselae and C. monstrosum from the Philippines and Palau interdigitated and were also very close to C. cardissa from Sulawesi (Fig. 2-8). In contra st, a specimen from PNG identified as C. cardissa differed by >5% in COI implying that morphology does not correspond with mtDNA boundaries in this clade. Within the genus Lunulicardia , two species that correspond to L. retusa and L. hemicardia were resolved (Figs. 2-5 & 2-6), and the inclusion of additional samples of the former species revealed cons iderable structuring within this lineage (e.g., between individuals from Tanzania and We stern Australia, Fig. 2-7).
38 Figure 2-7 COI phylogram (NJ, GTR model) of speciesÂ’ boundaries and geographic substructuring in Fragum (pale blocking) and Lunulicardia (dark blocking). Numbers following names are sample si zes and bootstrap values (1000 reps) where >90% are indicated below branches.
39 Figure 2-8 COI phylogram (NJ, GTR m odel) of speciesÂ’ boundaries in Corculum. Bootstrap values (1000 reps ) >90% are indicated be low branches. Blocking delineates similar geographic locales. Discussion Monophyly of the Cardiid Subfamily Fraginae? These results support the original Fragin ae circumscribed by Stewart (1930) and are in almost complete agreement with th e subfamilial delineation of Keen (1980) and Voskuil and Onverwagt (1989) (see Table 2-1). Circumscriptions differ in that both Keen (1980) and Voskuil and Onverwag t (1989) included the genus Afrocardium in the Fraginae (excluded from the subfamily in these analyses) and Ctenocardia victor is best supported as a distantly related cardiid by this study and not a fragine (Figs. 2-3 through 2-6). The present results contrast signifi cantly with the Fraginae as circumscribed by Schneider (1998), largely because of the high le vels of sequence divergence that resulted in very different placement of Parvicardium sensu lato representatives. These analyses do
40 not support the inclusion of divergent P. minimum and Papillicardium papillosum (Clade IV) in the Fraginae. That Clade IV is distan tly related to all tested fragines is supported by the recovery of one of the tested outgroup clades as more closely related to fragine lineages (Clades I-III) than P. minimum + Papillicardium papillosum (Fig. 2-6). However, although it is apparent that Clade IV is distantly relate d to the originally circumscribed subfamily (Clades II and III), it is equivocal whether the remainder of sampled Parvicardium (Clade I) will be well-supported subfamilial members. There is weak support for a clade comprised of some Parvicardium species, including P. exiguum (Clade 1), Ctenocardia + Trigoniocardia (Clade II) and Fragum , Lunulicardia and Corculum (Clade III), but this node is only recovered in the Bayesian analysis of the combined dataset and support is low (posterior pr obability=54, Fig. 2-6). Excluding Parvicardium entirely from the Fraginae was s upported by a recent large-scale molecular analysis of the Bivalvia. Parvicardium exiguum was recovered as sister to Fulvia mutica + Vasticardium flavum which was in turn, sister to three IWP fragines in the genera Corculum + Fragum (Giribet and Distel 2003). However, the inclusion of additional species of Parvicardium but, more importantly, appropriate molecular markers for resolution at deeper phylogenetic levels is necessary before circumscribing a new Fraginae. The lack of a perisiphonal suture results in confluence of the incurrent siphonal aperture and pedal gape, and is the str ongest of the morphologi cal synapomorphies uniting the Fraginae. As stated firmly by Sc hneider (1998: 326), Â“all fragines and only fragines lack a persiphonal sutureÂ”. However, live fragines effectively have a separate incurrent aperture, as they hol d the two mantle edges together muscularly at the ventral
41 margin of the incurrent aperture. Similar sepa ration of incurrent or excurrent apertures is common in many lineages of bivalves (e.g., mytilids, thyasirids: Bernard 1972, Payne and Allen 1991, respectively). The absence of mantle fusion at the ventral margin of the incurrent aperture is clearly secondary in fragin es, as it is present in all other cardiids, as well as almost all members of the Heterodonta, the order to which cardiids belong. Only a few other heterodonts lack mantle fusion around the incurre nt or excurrent apertures, most notably members of the Galeommatoidea. The absence of such mantle fusion in galeommatoids was thought to be plesio morphic, but new work suggests that galeommatoids may be secondarily simplif ied from higher heterodonts (Giribet and Wheeler 2002), implying that they may also have lost mantle fusion. Loss of mantle fusion may be related to mini aturization in these tiny cl ams, a hypothesis that also resonates in the Fraginae. Another morphological feature of fragines is a decrease in the complexity of their digestive sy stem (e.g., gut simplification in Trigonicardia and Apiocardia from Type V to Type IV, reductions in crystalline style and style sac in Corculum and the loss of ridges on the labial palps of Microfragum ), a trend attributed by Schneider (1998) to acquisition of photosym biosis, but negated by the aposymbiotic condition of the majority of fragine genera (Persselin 1998). An alternative hypothesis for digestive simplification is that it is also indicative of a history of miniaturization in fragine ancestry. Indeed, some of the Fraginae, including Parvicardium and some Fragum species, are among the smallest cardiids. The Early Diverging Fragine genus Parvicardium : Clades I and IV Schneider (1998) recognized the distinctiveness of Papillicardium and recovered a paraphyletic Parvicardium . Generic support of Papillicardium was novel and based on three morphological synapomorphies of Parvicardium that Papillicardium did not
42 exhibit: 1) distinct positioning of dorsa lmost siphonal tentacles to beyond the top adductor muscles, 2) single egg attachment by a double mucous membrane and 3) a trigonal shell shape (Schneid er 1998). This placement contrasted with earlier works by Kafanov and Popov (1977), Keen (1980), Vos kuil and Onverwagt (1989, 1991) and Van Aartsen and Goud (2000) who considered Papillicardium to be a subgenus of Parvicardium . The findings presented here support Papillicardium as distinct from all other shallow-water Mediterranean Parvicardium sampled in this study (Clade I). Although Papillicardium papillosum is recovered as sister to Parvicardium minimum, this may be due to long-branch attraction; bot h taxa were highly divergent for all gene regions sequenced. P. minimum is a deep-water species and the sample included from this study was from Sweden, geogra phically disjunct from other small Parvicardium and P. papillosum , which were hand-collected from the Mediterranean (Spain+France) (Table 2-3). Clade I is composed of three recognized species and one highly differentiated taxon (based on a single individual from La Herra , Spain) (Figs. 2-3 through 2-6). As P. exiguum, considered the type species in the most recent and comprehensive review of the Mediterranean Parvicardium (van Aartsen and Goud 2000, but also Hylleberg 2004 and others), is a member of this clade, the nomen Parvicardium should be restricted to this group. These preliminary results suggest that sm all European cardiids, presently allied to Parvicardium sensu lato, are highly divergen t lineages. Given these findings, geographically disjunct but presently unsampled Parvicardium species ( P. turtoni from South Africa and P. pinnulatum from the west Atlantic) are not expected to fall within this clade, which is currently restricted to morphologically-similar species from relatively
43 shallow Mediterranean habitats (Table 2-2). However, although P. turtoni is considered a member of the genus Parvicardium , the close morphological resemblance between P. turtoni and Papillicardium papillosum suggests that these two may be sisters (Kilburn and Rippey 1982, Plate 40:7). The equivocal findings generated by this study call attention to the need for a comprehensive sy stematic revision of the earliest diverging fragine genera, Parvicardium and Papillicardium . The Â“ CtenocardiaÂ” and Â“ TrigoniocardiaÂ” groups: Clade II The genera Trigoniocardia and Ctenocardia were recovered as a well-supported clade in Bayesian analyses of both mitochondrial and combin ed datasets (Figs. 2-4 & 25). This contrasts with Schneider (1998), who split these genera into two divergent fragine subclades (Fig. 2-2D) and recovered the Trigoniocardia group as sister to all fragines (other than Parvicardium sensu lato). The genus Americardia is most morphologically similar to Ctenocardia not Trigoniocardia, as recognized recently by Schneider (1998) as well as Vosku il and Onverwagt (1989), who synonymized Americardia with Ctenocardia because of phenetic similarity. However, in contrast with Schneider (1998) and Voskuil and Onverwagt (1989), but consistent with the present analyses, is the bulk of earlier taxonomic work (Stewart 1930, Clench and Smith 1944, Keen 1951, 1980, Olsson 1961 and Popov 1977) that recognized Americardia as a subgenus of Trigoniocardia . The present molecular phylogenetic findings, along with these older works, suggest that Clade II may be composed of separate IWP (CtenocardiaMicrofragum ) and American (Americardia-Trigoniocardia) radiations. This pattern is encountered again at lower taxonomic levels w ithin American lineages of Clade II (Figs. 2-5 & 2-6). Combined and mitochondrial data sets support a subclade of morphologically divergent East Pacific species: Trigoniocardia granifera, Apiocardia obovale and
44 Americardia biangulata to the exclusion of Americardia media from the Caribbean (Fig. 2-9). These findings resonate with recent studies of Caribbean reef corals (Fukami et al. 2004) that recovered multiregional genera as polyphyletic, thus supporting intraregional morphological radiation and interregional c onvergence. However, the inclusion of Caribbean T. antillarum , morphologically similar to T. granifera from the East Pacific, will likely falsify the hypothesis of strict intraregional radiation in the Americas. Within the large Ctenocardia clade, C. victor falls consistently within either a clade of outgroup species or in a deep unresol ved polytomy, clearly not closely related to other Ctenocardia (Figs. 2-3-2-6). The mo rphological disparity of C. victor, relative to other Ctenocardia, has long been appreciated. Wils on and Stevenson (1977) did not support placement of C. victor in Ctenocardia ; instead this species was allied to the genus Â“ Cardium Â” because of significant differences in hinge morphology relative to other conspecifics ( C. victor has a single right posterior la teral tooth, whereas all other Ctenocardia species have two posterior lateral teeth). The high levels of sequence divergence in all four gene regions reported here for multiple individuals, coupled with the morphological disparity observed by others , confirm the distin ctiveness of this species. A new genus will be erected for C. victor in a forthcoming publication revising Ctenocardia and allies. Relationships among the IWP genera Fragum , Corculum and Lunulicardia : Clade III Clade III includes the bulk of sha llow-water IWP species and all known photosymbiotic fragines, and is com posed of members of the genera Fragum , Lunulicardia and Corculum . In all analyses, Fragum is strongly supported as paraphyletic and in combined and mitochondrial datasets is composed of three subclades (Figs. 2-5 & 2-6). Schneider (1998 ) was uncertain of the position of Fragum (Fig. 2-2
45 D) and placed this genus in a trichotomy with Corculum + Lunulicardia and Microfragum + Ctenocardia + Americardia (1998:354) or as sist er to a clade of Corculum + Lunulicardia . Persselin (1998) favored the la tter topology and suggested that Fragum was likely paraphyletic, as supported by the results of this study. The taxonomic affinities of F. erugatum and F. sueziense have been long debated, with both species allied to no fewer than five different ge nera (see Hylleberg 2004). The results of this study unite these two as sisters, sister to all other tested Fragum, Lunulicardia and Corculum species (Figs. 2-5 & 2-6) . F. erugatum is endemic to SharkÂ’s Bay, Western Australia, where it is the dominant faunal component in many of the hypersaline reaches (e.g., Shell Beach). It is a morphologically variable species, with conspecifics exhibiting differences in shell shape, dentition and f eatures of the hinge often recovered among different classes of bivalves (LK pers. obs.& unpubl. data).
46 Figure 2-9 Geographic signature of majo r Fraginae groups overlaid on combined phylogeny. In contrast, F. sueziense is more morphologica lly conservative than F. erugatum , but more widespread geographically (e.g., througho ut IWP based on collections made in this study). F. sueziense, like F. erugatum , occupies a unique environment compared to all other known Fragum species; F. sueziense is entirely restricted to relatively turbid, subtidal environments typical of lagoons a nd large bays throughout its range. Like all other tested fragine species, F . aff . mundum from Tuvalu is reciprocally monophyletic. Initially thought to be a variant of F. mundum following preliminary morphological examination, F. aff. mundum was recovered as a member of the 25-rib clade and sister to F. carinatum (Figs. 2-3 through 2-6). Fi ne-scale habita t differences between F . aff. mundum and F. mundum provide a good means of differentiating between
47 these two entities in the field. The former i nhabits protected inner reef habitats, very different from the highly exposed out er reef microhab itat typical of F. mundum . Formal description, including discu ssion of diagnostic yet subtle morphological characters evident upon microscropical examination, as well as ecological preferences, is currently underway. Preliminary findings support Corculum cardissa as a single species, with morphological variability better explained in the context of ontogeny or local adaptation to environmental conditions (K awaguti 1950) (Fig. 2-8). However, the single divergent sample from Loloata Island in southern PNG hints at allopatric differentiation in the genus. The biology of Corculum supports allopatric speciatio n as well, given that studied individuals possess a short larval stage that would hinder, if not pr event, regular longdistance dispersal. Tests of allopatric species formation would be best tested via inclusion of Corculum from remote insular localities, including C. dionaeum from central Pacific islands as well as an Indian-ocean morph (Bartsch 1947). Moreover, if length of the larval stage is similar to those previously documented in Corculum (Kawaguti 1950), fine-scale differentiation among morphologica lly similar insular populations would also be expected. Geographic structuring was evident among all tested Fragum species (see Figs. 2-7 & 2-8) targeted for intensive population-leve l sampling. For example, individuals from Okinawa of F. loochoanum, F. unedo and F. mundum were consistently recovered as well-supported, reciprocally monophyletic lin eages compared with populations sampled from all other locales. However, relati vely large sample sizes (e.g., at least 10 individuals/locale) are gene rally necessary to conclusi vely establish reciprocal
48 monophyly (Kirkendale and Meyer 2004). As de monstrated in this study, reciprocal monophyly breaks down with increased population-level sampling of F. loochoanum from Tuvalu and F. sueziense from Papua New Guinea. Preliminary evidence of finescale geographic structuring, coupled with poor dispersal capabilities (e.g., for tested Corculum from Palau, Kawaguti 1950), suggests that this group is a good bivalve candidate for future phylogeographic initiati ves in the insular IWP (Paulay and Meyer 2002, Kirkendale and Meyer 2004, Meyer et al. 2005.
49 CHAPTER 3 CHARACTER TRAIT EVOLUTION Introduction Symbioses, loosely defined as one organi sm living in intimate association with another (Cowen 1983, Futuyma 1998), are a perv asive and enduring theme in the history of life (Margulis and Fester 1991, Margulis 1998, but see Gray et al. 1999). Extant partnerships may impart members with nove l capabilities often referred to as key innovations and are widely dist ributed phylogenetically with re presentation in all three Â“domainsÂ” (e.g., Eukarya, Archaea, Eubacter ia; Norris 1986, Beckage 1998, Bourtzis and O'Neill 1998, McFall-Ngai 1998, Meeks 1998, Moran and Telang 1998, Richardson 2001). The regular discovery of new sy mbiotic associations (e.g., shelled gastropods/endozoic algae (Ber ner et al. 1986), sponge/cya nobacteria (Thacker and Starnes 2003), and marine isopods/cyanobacteria (Lindquist et al. 2005)) at increasingly fine-scales (e.g., symbiont genes in host genom es (Mujer et al. 1996)), coupled with the impact that symbionts can have on host reproduction and fitness (e.g., Wolbachia (Telschow et al. 2005)), atte sts to their functional importance as well as to the evolutionary insight offered by their study. Symbioses are considered facultative or oblig ate, with the latter most often typified by nutrient exchange between partners. When acr oss trophic levels, these associations are referred to as Â“biotrophic symbiosesÂ” (Smith & Smith 1990, Jumpponen & Trappe 1998a). One type of biotrophic symbiosis, chem osymbiosis, forms the basis for one of the most recently discovered deep-sea faunas, the hot vent community (Corliss 1979, Distel
50 1998). A diverse assemblage of marine inverteb rates serve as hosts in these associations, including bivalve and gastropod molluscs, vestimentiferan, pogonophoran, as well as nematode worms. Among bivalves, chemosymbiosis has been found to be common both phylogenetically and ecologically (Reid 1990, Distel 1998); not less than 5 bivalve families include chemosymbiotic members (Williams et al. 2004). The best studied, oldest and most diverse of th ese is the Lucinidae, with approximately 50 living genera widely distributed from the deep-sea (where species are often massive; see Bouchet and von Cosel 2004) to shallow-water sandy habita ts adjacent to coral reefs (Oliver 1986, Glover and Taylor 1997, 2001; Taylor and Glover 1997a, b, 2002). Morphological adaptations for bivalve chemosymbiosis can be extensive and may include: 1) reductions in the gut (as it is little used), 2) cooption and enlargement (also th ickening) of the gills for storage and processing of endosymbiotic bacteria and 3) a long and vermiform foot that is used to probe the sediments for su lphur-rich layers (Rei d 1990, Distel 1998, Taylor and Glover 2000, Williams et al. 2004). Chemosymbiotic status and putative adap tations among bivalves have been well documented (e.g., Reid 1990, Distel 1998, Tayl or and Glover 2000, Dufour and Felbeck 2004) and support multiple, independent orig ins of chemosymbiosis among distantly related bivalve lineages. However, the exact number of independent origins is still being worked out, as new research suggests that some chemosymbiotic families that were thought to be sisters (e.g., Lucinidae, Thyasi ridae) are not (Williams et al. 2004). These findings provide insight into the evolution of complex characters, suggesting that some
51 are less subject to phylogenetic inertia than previously thought (e.g., mode of development in calyptraeid gastropods, Collin 2004). In contrast to the relatively recent di scovery of chemosymbiotic partnerships, photosymbioses, pervasive in shallow-water tr opical environments, have been known for centuries. A description of the associati on between corals and photosymbionts (Dana 1846) was followed by description of photos ymbioses in giant clams (Cardiidae: Tridacninae, Yonge 1936). Just five year s later, Kawaguti (1941, 1950) described zooxanthellate photosymbionts in a s econd, but much smaller bivalve, Corculum cardissa (Cardiidae: Fraginae). Since accounts in C. cardissa , several other frag ine species have been found to possess photosymbionts: F. fragum and F. unedo (Kawaguti 1983 and Umeshita & Yamasu 1985, Yamasu 1988 a, b), F. loochoanum (Ohno et al.1995), Lunulicardia retusa (Schneider and Carter 2001), F. erugatum (Morton 2000) and L. sp. 1, F. mundum, F. nivale, F. sueziense and F . sp. 11 (Persselin 1998). Based on observed digestive system simplifications in generic exemplars, coupled with the then ubiquitous occurrence of photosymbiosis in tested fragines, Schneid er (1998) proposed that all members of the subfamily Fraginae were lik ely photosymbiotic. In contrast, Persselin (1998) was more conservative and suggested that photosymbiosis was restricted to Fragum , Corculum and Lunulicardia , but absent from most other fragine lineages. Most photosymbiotic species, including the hundreds of sc leractinian coral species and other cnidarians, as well as didemnid ascidians and opisthobranch mollusks, have large areas of photosymbiont-p acked tissue exposed to li ght (Cowen 1983, LK pers. obs.). In contrast, the thick, opaque shells and infaunal habit of bivalves are not conducive to high light exposure optimal for photosymbionts, yet many species are
52 photosymbiotic. Given these obviou s constraints, how do these bivalves facilitate light penetration to photosymbi onts? The giant clams ( Tridacna and Hippopus ) possess a similar suite of morphological a nd behavioral features that appear to be adaptations for a symbiotic lifestyle (Rosewater 1965, Yonge 1936, 1981, Schneider and " Foighil 1999). In giant clams, the posterior portion of the animal that lies upward has been greatly expanded at the expense of its anterior anatom y. The mantle of this posterior region has greatly hypertrophied and is expos ed to light between the valves that are habitually kept widely agape, and, in Tridacna (but not Hippopus ), also extend beyond the shell margin. Unlike most other cardiids or heterodont biva lves, tridacnines are epibenthic and thus have moved out from the sediment into the s unshine. It is widely accepted that the added nutrition provided by photosymbionts (Trench et al. 1981, Fitt and Tr ench 1981, Fitt et al. 1986, Fitt et al. 1993), coupled with a fully f unctional gut, has allowed giant clams to attain a large size compared to most other bivalves (adults always >10 cm, with the largest, T. gigas, up to 1 m across). The term adaptation used here follows Gould and Vrba (1982) to describe a feat ure that has evolved via natu ral selection for its present function. Two other terms are pertinent to this study: 1) exaptation refe rs to features that were co-opted for their present adaptive func tion and 2) aptation describes situations where the evolutionary context is unclear. Whereas all extant Tridacninae are photos ymbiotic and exhibit a similar suite of adaptations, the closely related cardiid subfam ily Fraginae has traveled a much different path. Not all fragines are photosymbiotic, but th ose that are exhibit a striking diversity of aptations to enhance light exposure to zooxanthellae (Kawagu ti 1950, Ohno et al. 1995, Carter and Schneider 1997, Persseli n 1988, Morton 2000, Schneider and Carter
53 2001)(Fig. 3-1). This is best illustrated by contrasting two of the best-studied photosymbiotic fragines, Fragum unedo and Corculum cardissa . The photosymbiotic aptations of F. unedo are reminscent of giant clams, with individuals gaping their valves and splaying their hypertrophied mantles onto the sediment surface (Ohno et al. 1995). A very different strategy typifies C . cardissa , where a thin and flatte ned shell, an epifaunal existence and translucent posteri or shell regions are most sugge stive of a solar panel-like habit (Ohno et al. 1995). The dive rsity of putative adaptations in fragines contrasts with the uniformity of response encountered in photosymbiotic tridacnines (Yonge 1980) and chemosymbiotic lucinoids (Taylor and Glover 2000). A lthough several morphological aptations within photosymbiotic fragines appear parallel, the fact th at others are widely divergent provides evidence for the importance of historical contingency in the evolution of form (Gould and Lewontin 1978, Ohno et al. 1995). Although the bulk of documented photosymbi otic clams comes from these two cardiid subfamilies, five other species of bivalves have been found to host symbiotic algae. Fluviolanatus subtorta , a bizarre clam from shallow, tropical Australian seas tentatively assigned to the Trap ezidae, has been reported to harbor zooxanthellae (Morton 1982). Zoochlorellae have been reported fr om several bivalve species including the freshwater Anodonta cyngea and Unio pictorum (Pardy 1980), the marine giant scallop Placopecten magellanicus (Naidu & South 1970), and the temperate marine cardiid Clinocardium nuttalli (Hartman and Pratt 1976, Jones and Jacobs 1992). However, in these associations the relationship is typica lly facultative and the bivalves are little modified morphologically (although see Morton 1982 for F. subtorta and Pardy 1980 and Farmer et al. 2001 for Anodonta ).
54 Figure 3-1 Schematic diagram of some shell adaptations in photosymbiotic Tridacninae and Fraginae. A. Tridacna squamosa, B. Fragum unedo , C. Fragum loochoanum , D. Fragum fragum , E. Lunulicardia hemicardia and F . Corculum cardissa . Dark circles denote the ligament, dashed line indicates sediment surface, arrows indicate incide nt light and hatched circles represent windows.
55 This is in contrast to the apparently obligate and intimate association with zooxanthellae confirmed for fragines and tr idacnines (Farmer et al. 2001, Norton et al. 1992), where nutritional exchange and/or abundant morphological and behavioral aptations have long been known (e.g., Ka waguti 1950, Yonge 1980, Trench et al. 1981, Fitt and Trench 1981, Fitt et al. 1986, Wats on and Signor 1986, Janssen 1991, Fitt et al. 1993, Ohno et al. 1995, Carter and Schneider 1997, Persselin 1998). Photosymbiotic status (Kawaguti 1950, O hno et al. 1995, Schneider 1998, Persselin 1998, Morton 2000), and putative morphological adaptations for photosymbiosis in fragines (Persselin 1998, Ohno et al. 1995, Carter and Schneider 1997, Schneider and Carter 2001) have been studied and docume nted for decades. However, comprehensive assessment of photosymbiotic status and formal scoring of aptive morphologies, behavior and habitat is lacking. Even given a welldelineated character matrix of putative morphological aptations, coupled with a thorough understanding of photosymbiotic status, it is impossible to place this informa tion in an evolutionary context without a phylogeny. A comprehensive hypothesis of the relationships among subfamily members is necessary to pinpoint the origin of photos ymbiosis in the group, and to begin tracking the evolution of morphologica l character traits hypothesized as aptive and adaptive for photosymbiosis. The specific aims of this study were thr eefold. First, as many Fraginae species as possible were collected and examined for sy mbiotic algae, to evaluate the taxonomic distribution of the photosymbiotic condition. Sec ond, character traits judged to be potent aptations for a photosymbiotic lifestyle were compiled from the literature and character states for these traits were formally scor ed from both literature accounts and museum
56 specimens for a wide range of both photosym biotic and non-photosymbiotic cardiids. Third, these two bodies of information were then placed in a phylogenetic context in order to understand the origin of photosymbi osis and the evolution of photosymbiotic aptations in the group. The fairly comprehens ive Fraginae phylogeny available permitted this latter initiative (Chapter 1). Materials and Methods Specimen Acquisition Animals were collected by hand or dr edging, and supplied by collaborators. Samples were fixed in ethanol for molecular analyses and, where possible, also in 5% seawater formalin for morphological study. Most studied samples (shells and tissues) are deposited in the Florida Museum of Natural History (Table 3-1). Determination of Photosymbiotic Condition Assessment of photosymbiotic status for sampled Fraginae was undertaken by examining the mantle, gill and foot of liv e collected animals in the field, under a microscope when available. A small piece of tissue was placed on a glass slide to check for zooxanthellar photosymbionts, the photos ymbiont most common to shallow water IWP corals and giant clams. Zooxanthellae have a characteristic shape (completely spherical), color (dark or golden brown) and size (approximately 5-8 Âµm), cells that fit this description were considered zooxan thellae. The taxonomic identity of fragine symbionts was not directly determined (e.g., by genetic analyses). However, because most, if not all photosymbiotic zooxanth ellae are dinoflagellates of the genus Symbiodinium, it is very likely that they all pertain to this algal genus. Live animals were photographed unrelaxed as well as relaxed (with magnesium chloride or Epsom salts) when possible. When a microscope was not av ailable in the field, the color of animals
57 Table 3-1 Species, distribution and vouchers of sampled Fraginae (asterisks denote type species of genera, see Chapter 1 for aut horities). Vouchers are representative samples used in morphological analyses , contact LAK for a complete list of sampled individuals. Species Distribution Vouchers Fragum fragum * IWP UF299282 Fragum scruposum IWP UF374114 Fragum loochoanum IWP UF299448 Fragum carinatum IWP UF351691 Fragum mundum IWP UF296894 Fragum aff . mundum IWP UF348016 Fragum unedo IWP UF299291 Fragum erugatum IWP UF299293 Fragum sueziense IWP UF299263 Lunulicardia retusa IWP UF291497 Lunulicardia hemicardia IWP UF299269 Corculum cardissa* IWP UF286449 Trigoniocardia granifera* East Pacific UF359687 Apiocardia obovale* East Pacific UF351671 Americardia biangulata East Pacific UF351615 Americardia media* West Atlantic UF347556 Ctenocardia fornicata IWP UF286471 Ctenocardia victor IWP UF288935 Ctenocardia gustavi IWP UF351689 Microfragum festivum* IWP UF374119 Parvicardium exiguum* East Atlantic UF374120 Parvicardium vroomi East Atlantic UF374116 Papillicardium papillosum East Atlantic UF374115
58 was assessed to see whether they had th e dark, brownish tissue discoloration characteristic of photosymbiotic fragines. Th ese animals were then fixed in formalin for later, lab-based, microscopic confirmati on of symbiont occurrence. Tentative photosymbiotic status of speci mens supplied by collectors ( Parvicardium exiguum by J. Vidal and Apiocardia obovale by R. Collin) was established via discussion, as ethanol preservation of these specimens precluded photosymbiotic verification. Morphological Character Trait Analyses Eight morphological and ecological characte rs, many previously considered to be photosymbiotic adaptations (Kawaguti 1941, 1950, Watson and Signor 1986, Janssen 1991, Ohno et al. 1995, Carter and Schneider 1997, Persselin 1998, Schneider 1998), were scored from 11 non-photosymbiotic a nd 12 photosymbiotic frag ines for character trait reconstruction (Table 3-1, bolded species). Measurem ent or assessment of: 1) window shell microstructure (WSM), 2) decreas ed pigmentation in posterior shell region or Â“pigment windowsÂ” (PIGW), 3) maximum observed adult height, the traditionally scored estimate of size in bi valves (SIZE)(Fig. 3-2A), 4) habitat (HAB), 5) exposure (EXP), 6) mean keel angle (KEEL)(Fig. 32C), 7) mean posterior shell flattening (FLAT)(Fig.3-2B) and 8) mean linearized surface area to volume ratios (SAV)(Fig.3-2D) was scored for adult individuals of each species. Below I define and discuss each character, and evaluate their adaptive status. Character Data Windows Window shell microstructure (WSM) is comp lex but is created by: 1) thinning of the outer fibrous prismatic (FP) layer, 2) in cursion of the FP outer shell layer into middle shell regions and 3) passing of fibrous prisms into less translucent shell microstructural
59 layers (e.g., dissected crossed prismatic (DCP) that extend these Â“windowsÂ” toward the interior of the shell, Seilacher 1990, Carter and Schneider 1997, Persselin 1998, Schneider and Carter 2001). In itially thought to function as a fiber optic system to disperse light to the zooxanthellae situated within the mantle of Corculum cardissa (Seilacher 1990), later studies revealed that the shell micr ostructural adaptations more likely function as triangular light-condensing lenses, consis tent with the often deep zooxanthellar placement in the mantle of C. cardissa (Carter and Schneider 1997). Because this feature results in increased sh ell translucency, and has only been reported from photosymbiotic fragines to date, it is widely considered an adaptation for photosymbiosis (Gould and Lewontin 1978, Watson and Signor 1986, Seilacher 1990, Carter and Schneider 1997). Reduced shell pigmentation also increase s shell translucency similar to window shell microstructure, however th is can occur with or without microstructural alteration (Watson and Signor 1986, Carter and Sc hneider 1997, Persselin 1998). Like window shell microstructure, decreas ed shell pigmentation is mo st commonly observed in the posterior shell region that is incident to light. Because this feature increases shell transclucency, and has only been reported from photosymbiotic fragines to date, it is also considered an adaptation for photosymbiosis . Both microstructural and pigmentation windows were scored as absent (0), presen t (1) or uncertain (2). Presence of microstructural and pigmentation windows wa s obtained almost exclusively from the literature (Carter and Schne ider 1997, Persselin 1998).
60 Figure 3-2 Morphometrics used in this study to quantify putative photosymbiotic attributes. Dotted lines indicate measured distances. A. Height (H; cm), B. Angle A (in degrees), the angle formed by the posterior shell margin Â“crestÂ” and out to the keel of the left and right valve (to capture posterior shell flattening), C. Angle B (in degrees), the angle formed by the strongest keel rib to 4 ribs on either side (to capture the angle of the keel), D. Posterior surface area incident to light (shaded) measur ed for surface area to volume ratios.
61 However, the absence of window shell micr ostructure and/or reduced pigmentation is quite apparent from visual inspection of shells. Valves were examined under a microscope with a light source held incident to the posterior shell surface to establish the absence of small (approximately 1mm2, althoug h even smaller regions were common in smaller specimens) zones of translucency, th at when present are interspersed fairly regularly with more opaque shell regions. Moreover, comparison with a number of available windowed species (e.g., F . mundum , C. cardissa and F. nivale ) aided in scoring absence in previously unexamined species. An alyses similar to those performed by Carter and Schneider (1997), Persselin (1998) and Schneider and Carter (2001) need to be conducted for confirmation of preliminary observations reported here. Maximum observed adult size Small size results in thinner shells that lead to increased shell translucency. This in turn, increases light penetration to the she ll interior and is considered aptive for a photosymbiotic lifestyle. Moreover, small si ze also yields elevated surface area to volume ratios relative to larger shells of similar shape. Becau se increased surface area to volume ratios increase light exposure to photos ymbionts, this character state is also considered aptive for a photosymbiotic lifestyle. Size was based on extensive examination of several museum collections (w et and dry material from Florida Museum of Natural History, University of Florida bu t also University of Michigan Museum of Zoology, Museum National dÂ’Histoire Naturelle , Paris and the Museum of Comparative Zoology, Harvard), as well as new material collected during this study. Certain species were better represented than others in collections, and this influenced observed size data. Maximum observed adult size was scored from consideration of at least 10 specimens for each examined species (with hundreds obs erved for common species such as F. fragum ),
62 except F. aff . mundum (4 individuals), A. biangulata (2 individuals), A. obovale (3 individuals) and C. gustavi (5 individuals) . The largest individual per species was then chosen and recorded, and these data were broken down into eight, 1-cm increments: 1) 0.1-1 cm through 8) 7.1-8 cm (Fig. 3-2A and Appendix A). Habitat Habitat and microhabitat were chosen to ch aracterize ecology in the Fraginae. Five habitat categories: 1) in ner reef flat, 2) outer reef flat, 3) lagoon, 4) reef slope and 5) nonreef were recognized and scored for each spec ies. Habitat captures depth (see Fig. 3-3 for reef-associated habitats), with reef flats the shallowest environment (0-2m), generally providing the highest light leve ls compared with all other reefal habitats (e.g., deeper lagoon and reef slope environments, >5 m)(F ig. 3-3). The shallowest depths, found in reef flat habitats, offer the highest light environments and are considered aptive for a photosymbiotic lifestyle. Habitat records, made during the duration of this study and throughout large expanses of sampled speciesÂ’ geographic ranges, were the primary data source (now housed at the Florida Museum of Natural History, see LAK for full records). If not hand-collected during th e duration of this study, habi tats were based on museum records (most notably the extensive material at the Museum National dÂ’Histoire Naturelle, Paris) from live-collected sp ecimens. Although most species are habitatspecific, those that were not were scored accordingly. This strategy was not followed where large population sizes of a given speci es were overwhelmingly from one habitat type, with a few outliers f ound in adjacent habitats (e.g., F . fragum most commonly occupies inner reef flats but was also taken rarely from adjacent shallow lagoons).
63 Figure 3-3 Habitat of reef-asso ciated Fraginae included in th is study. Stippling denotes sand.
64 The second ecological character scored was microhabitat. High levels of exposure, typical of an epifaunal existence, allow ma ximal light penetration to zooxanthellae for photosymbiotic taxa compared with a less exposed, infaunal existence. An exposed, epifaunal existence is unique to photosymbiotic fragines and is thus considered adaptive. To characterize exposure in fragines, two cat egories were scored: 1) infaunal and 2) epifaunal. The number of specimens examined for bot h ecological characters was similar to that described for maximum adult size, with at least 10 live-collected specimens surveyed per species, except for F . aff. mundum (4 individuals), A. biangulata (2 individuals), A. obovale (3 individuals) and C. gustavi (5 individuals) (Appendix A). Shell measurements The presence of a pronounced keel is considered aptive for a photosymbiotic lifestyle, as a strong keel minimizes burial depth; a sh allow infaunal existence permits increased levels of irradiance to the posterio r shell surface compared with a more deeply buried existence. Photosymbiotic species were expected to exhibit more pronounced keels (considered aptive) than non-photosymbio tic species. Keel angle was measured as the angle formed at the shell margin by th e most pronounced keel rib (the angle was measured from the keel rib to the 4th rib on either side), with the smallest angle describing the sharpest keel (Fig. 3-2C). M ean values were calculated for each species and data were then split into five, 20Â° incr ements: 1) 40-60Â° through 5) 120-140Â°. Closely related to keel angle is posterior shell fl attening, which also decr eases burial depth and allows for increased levels of irradiance to the posterior shell regi on relative to light penetration possible with deeper burial. Photosymbiotic species were expected to exhibit more pronounced posterior shell flattening (c onsidered aptive) than non-photosymbiotic
65 species. Posterior shell flattening was measured as the angle formed by the posterior crest (highest point of the posterior shell slope) out to the keel on each valve, with the largest angle describing the flattest pos terior shell area (Fig. 3-2B). Mean values (per species) were broken down into seven, 10Â° increments beginning at 140-130Â° and ending at 9080Â°. Surface area to volume ratio is predicte d to be elevated in photosymbiotic compared with non-photosymbiotic species, as this increases the lig ht gathering surface incident to light. Higher surf ace area to volume ratios were thus considered aptive for photosymbiotic species. In order to test this hypothesis, mean linearized surface area to volume ratios were calculated. Volume was estimated by measuring the amount of water (ÂµL) required to fill the two valves using a pipetter (Pipetteman brand). The posterior shell surface area was measured by: 1) phot ographing the posterior shell surface with a digital camera (Nikon Coolpics 3200 camera) a nd 2) estimating the surface area of this region (cm3) using ImageJ software (NIH freeware)(Fig. 3-2D). For photography, the posterior shell plane was set parallel to the surface of a table using cotton balls. Surface area to volume (cm3) ratios were linearized (cube r oot of volume and square root of surface area) to standardize for body size; these transforme d data were plotted to visualize trends. Means were calculated for each species and these data were grouped into six, 20 unit increments: 1) 0.10-12 through 6) 0.20-0.22 (Appendix B). Surface area to volume ratios, keel and shell flattening measurements were compiled from a wide range of cardiid species, including photosymbiotic and nonphotosymbiotic species of fragines and non-fragines to test the hypothesis that photosymbiotic species would exhibit elevated surface area to volume ratios, keel angle
66 and posterior shell flattening relative to non-photosymbiotic species (Figs.3-2BD)(Appendix C). These three measurements were tested to examine if they were significantly correlated. Anal yses of variance (single f actor ANOVA) were conducted on log-transformed data to test for significan t differences between 23 photosymbiotic and non-photosymbiotic species, with mean and 95% confidence intervals plotted to examine these trends visually (SAS, Excel). Principa l component analyses (PCA, SAS) were run to extract factors in the continuous dataset that explained variati on among cardiids and to further document differences among photosymbi otic and non-photosymbiotic species. Character Trait Evolution Characters were mapped onto a subset of the fragine phylogeny (Chapter 1) to understand the pattern of characte r evolution and reconstruct an cestral characte r states of the photosymbiotic clade (MacClade v. 4.08 OSX, Maddison and Maddison 2000). While fragine species were generally well resolved by molecular methods, species boundaries in Corculum were not. The status of some species , delimited by the onl y review (Bartsch 1947), was called into question by my molecular analyses. C. cardissa is used here for all Corculum taxa recognized by Bartsch (1947), except C. dionaeum . Results Photosymbiotic Status All tested representatives of 3 fragine genera ( Corculum, Lunulicardia and Fragum ) were entirely photosymbiotic, while no sampled members of the other 7 fragine genera ( Trigoniocardia, Apiocardia, Americar dia, Ctenocardia, Microfragum, Papillicardium and Parvicardium ) were found to host photosymbionts (Table 3-2).
67 Character Trait States Windows Although microstructural windows have been reported from four photosymbiotic fragine species thus far ( F. mundum, F. nivale, L. sp. 1 and C. cardissa , Carter and Schneider 1997, Persselin 1998), only half of these ( F. mundum and C. cardissa ) were included in the phylogenetic study (Table 3-3) . Evidence of intermediate microstructural window formation was reported for F. fragum (Persselin 1998). Pigment windows are presently documented from these same five species (Carter and Schneider 1997, Persselin 1998) (Table 3-3). Additionally, translucent pos terior shell regions were observed in all tested specimens of F . aff. mundum , and many smaller (e.g., juvenile) specimens of Lunulicardia hemicardia and L. retusa . However, it is uncertain whether these are due to microstructural alteration or decreased pigmentation, and these were scored as uncertain (Table 3-3). No other tested photosymbiotic or non-photosymbiotic fragines possessed transclucent areas in the posterior slope of their shells based on gross examination. Maximum observed adult size Photosymbiotic fragines in cluded several small species but also a wide diversity of sizes, running the continuum of sa mpled categories (Table 3-3). Three nonphotosymbiotic fragines, Papillicardium papillosum, Parvicardium vroomi and Microfragum festivum were also of the smallest size category. The two largest sampled fragine species, C. cardissa and F. unedo, were photosymbiotic.
68 Table 3-2 Photosymbiotic status of sampled Fraginae. Species Determination1-4 ~No. Photosymbionts Fragum fragum Field2 50 Present Fragum scruposum Field2 20 Present Fragum loochoanum Field2 20 Present Fragum carinatum Lab3 13 Present Fragum mundum Field2 3 Present Fragum aff. mundum Lab3 3 Present Fragum nivale Literature1 Present Fragum unedo Field2 5 Present Fragum erugatum Lab3 20 Present Fragum sueziense Field2 6 Present Fragum sp. 11 Literature1 Present Lunulicardia retusa Literature1 Present Lunulicardia hemicardia Lab3 2 Present Lunulicardia sp. 1 Literature1 Present Corculum cardissa Literature1 3 Present Trigoniocardia granifera Field2 10 Absent Apiocardia obovale Collaborator4 Absent Americardia biangulata Field2 2 Absent Americardia media Field2 3 Absent Ctenocardia fornicata Literature1 Absent Ctenocardia victor Literature1 Absent Ctenocardia gustavi Lab3 1 Absent Microfragum subfestivum Lab3 3 Absent Microfragum festivum Field2 10 Absent Parvicardium exiguum Collaborator4 Absent Parvicardium scriptum Field2 10 Absent Parvicardium vroomi Field2 10 Absent Papillicardium papillosum Field2 8 Absent 1Status previously known (see text for references). 2Microscopic examination of live animals. 3Microscopic examination of formalin-fixed animals. 4Microscopic examination of ethanol-fixed animals.
69 Table 3-3 Morphological character matrix for mapping traits of sampled Fraginae. Photosymbiotic species in bold. No Species WSM1 PIGW1 SIZE2 HAB3 EXP4 KEEL5 FLAT6 SAV7 1 Fragum fragum 2 1 5 1 1 3 3 3 2 Fragum scruposum 0 0 2 1 1 3 5 1 3 Fragum loochoanum 0 0 2 1 1 4 5 2 4 Fragum carinatum 0 0 1 1 1 4 4 2 5 Fragum aff mundum 2 2 1 1 1 3 2 2 6 Fragum mundum 1 1 1 2 2 3 2 2 7 Fragum unedo 0 0 8 1 1 4 3 3 8 Fragum erugatum 0 0 2 5 1 4 6 1 9 Fragum sueziense 0 0 1 3 1 5 5 2 10 Lunulicardia retusa 2 2 5 1 1 3 3 4 11 Lunulicardia hemicardia 2 2 5 1,5 1 3 2 3 12 Corculum cardissa 1 1 6 1,2 2 1 1 6 13 Americardia media 0 0 4 1,2,4 1 4 4 2 14 Americardia biangulata 0 0 2 5 1 3 5 2 15 Trigoniocardia granifera 0 0 2 5 1 3 6 2 16 Apiocardia obovale 0 0 2 5 1 3 4 2 17 Microfragum festivum 0 0 1 1,2,3 1 5 4 1 18 Ctenocardia fornicata 0 0 4 4 1 4 4 3 19 Ctenocardia gustavi 0 0 3 4 1 4 3 2 20 "Ctenocardia" victor 0 0 3 4 1 4 5 2 21 Papillicardium papillosum 0 0 1 5 1 5 4 1 22 Parvicardium vroomi 0 0 1 5 2 4 4 1 23 Parvicardium exiguum 0 0 2 5 2 4 3 1 10) absent, 1) present, 2) uncertain 2Maximum adult size (height in cm, Fig. 2A): 1) 0.1-1, 2) 1.1-2, 3) 2.1-3, 4) 3.1-4, 5) 4.1-5, 6) 5.1-6, 7) 6.1-7, 8) 7.1-8. 3Habitat: 1) protected, inner reef flat, 2) exposed, outer reef flat, 3) lagoonal, 4) reef slope, 5) non-reef 4Exposure: 1) shallowly buried/infaunal, 2) exposed/epifaunal 5Keel angle (Fig. 2C): 1) 40-60, 2) 60-80, 3) 80-100, 4) 100-120 and 5) 120-140 6Posterior shell flattening (Fig. 2B): 1) 140-130, 2) 130-120, 3) 120-110, 4) 110-100, 5) 100-90 and 6) 90-80 Habitat The most common habitat occupied by photos ymbiotic fragines was the reef flat (Table 3-3). However, two species occupi ed relatively divergen t habitats: 1) F.
70 erugatum, endemic to SharkÂ’s Bay, Wester n Australia, was the only photosymbiotic species from a non-reefal habitat, while 2) F. sueziense was restricted to turbid lagoonal settings. All but two photosymbiotic species, F. sueziense and L. retusa , were encountered in <2 m (several of the often tens to hundreds of specimens examined were encountered deeper, but are not considered typical of that depth) with only F. sueziense restricted to deeper depths (Table 3-3). Two tested photosymbiotic species, Corculum cardissa and F . mundum , and two non-photosymbiotic fragines, Parvicardium vroomi and Parvicardium exiguum , were epifaunal; all other species were infaunal. Shell measurements Significant differences were observed for single factor ANOVAs run among the three shell measurements (log keel angle, l og posterior shell flattening and log linearized surface area to volume) for twenty cardiid sp ecies (p<0.0001 for all pairwise comparisons with raw data, available from LAK) (Fig. 3-4). Mean and 95% confidence intervals revealed that photosymbiotic species genera lly exhibited the greatest range of observed values for all three measurements, but that the bulk of species overlapped with nonphotosymbiotic species (Fig. 3-4). This latter point is best illustrated by comparison of keel angle for tested species, where photosymbiotic species overlapped with nonphotosymbiotic species almo st entirely (Fig. 3-4A). Corculum cardissa was by far the most divergent morphotype and was significantl y different than all other sampled species in keel angle and surface area to volume, overl apping only with a few species in posterior shell flattening confidence intervals (Fig. 3-4A-C). All three shell measurements were significantly correlated (Pearson corre lation coefficients of p=0.0001 at =0.05 for all pairwise comparisons).
71 Increased taxon sampling for surface area to volume ratios reinforced: 1) that the bulk of photosymbiotic fragines were lit tle differentiated from non-photosymbiotic species and 2) that photosymbio tic species exhibited a wide range of shell forms (Fig. 35). The least differentiated forms, exhibiti ng the most overlap with non-photosymbiotic fragines and non-fragine cardiids, were F. erugatum, F. scruposum, F. loochoanum and F. sueziense , with C. cardissa, L. hemicardia, L. retusa, F. fragum , F. unedo and F. mundum at the other end of the morp hological continuum (Fig. 3-5). Principal component analyses were perf ormed on log-transformed mean surface area to volume ratios, mean posterior shell flattening, mean posteri or distance (as a proxy for posterior shell flattening) and mean keel angle for a subset of sampled species (see Appendices A & B). Eigenvalues greater than one, coupled with a cumulative variation of 87%, indicated that the ma jority of observed variation was due to the first two principal components (Fig. 3-6). These findi ngs indicate that bot h photosymbiotic and non-photosymbiotic species become larger and more crested to similar degrees (second principal component, Fig. 3-6). However, onl y photosymbiotic species make it above a Â“thresholdÂ” of approximately 1 for the first pr incipal component that described keel angle and posterior shell flattening (Fig. 3-6). This suggests that only photosymbiotic species become highly flattened and keeled, w ith this morphospace unoccupied by nonphotosymbiotic species.
72 Figure 3-4 Mean and 95% confidence intervals for three shell measurements (n=3+) for non-photosymbiotic and photosymbiotic ca rdiids. Photosymbiotic species are in yellow non-photosymbiotic species in black, with dotted lines denoting non-fragines. The y-axis units are lo g-transformed, surface area to volume ratios standardized for body size.
73 Figure 3-5 Significant differences among 33 species of cardiids, including photosymbiotic and non-photosymbiotic fragines for log-transformed surfacearea to volume ratios standardized fo r body size and plotted as means and 95% confidence intervals.
74 Character Trait Evolution A phylogenetic hypothesis of Fraginae rela tionships (Chapter 1) allowed mapping of character trait data to explore the evolution of photos ymbiosis in the subfamily. However, before character trait states are traced, relevant phylogenetic findings are briefly reviewed. Four major fragine clades were resolved : 1) Clade I composed of all but one species in the genus Parvicardium , 2) Clade II comprised all sampled representatives from the genera Trigoniocardia, Apiocardia, Microfragum , Americardia and Ctenocardia, except Ctenocardia victor , 3) Clade III comprised all species in the genera Fragum, Corculum and Lunulicardia and 4) Clade IV included two species, Parvicardium minimum and Papillicardium papillosum , distantly related to the other 3 clades (Fig. 3-4). Although poor resolution at deeper node s made drawing inferences concerning subfamilial monophyly difficult, lower levels were consistently well resolved (Fig. 3-7). Three subclades were recovered within Clade III: 1) an earliest diverging subclade ( F. sueziense, F. erugatum )(PSI), 2) a Â“25-ribÂ” group ( F. aff. mundum, F. carinatum, F. scruposum and F. loochanum )(PSII) and 3) a subclade uniting species from the genera Fragum, Lunulicardia and Corculum ( F. fragum, F. unedo, F. mundum, C. cardissa , L. retusa and L. hemicardia ) (PSIII)(Fig. 3-7). Fragum , as currently defined, was recovered as paraphyletic, giving rise to Corculum and Lunulicardia. A single well-supported origin of phot osymbiosis was supported with algal symbionts acquired at the base of Clade III ( Fragum, Lunulicardia and Corculum ) (Table 3-2, Fig. 3-7).
75 Figure 3-6 Principal component analysis for four morphological characters (Flat, Keel , Height and Crest) for photosymbiotic and non-photosymbiotic cardiids. Thes e two axes explain 87% of the va riation in the te sted dataset.
76 Figure 3-7 Origin of photosymbiosis trac ed onto Bayesian 50% majority topology for the combined dataset of Chapter 1. Posterior probabilities above branch es, maximum parsimony bootstrap values below branches for major groups. Dashed boxes denote PSI, PSII and PSIII, the th ree photosymbiotic subclades of Clade III ( Fragum , Lunulicardia and Corculum ).
77 With regards to character trait evolution, a given char acter state was considered aptive for photosymbiosis if it enhanced light penetration to photosym bionts, relative to other states.Although putative aptions were of ten derived from the literature (Kawaguti 1941, 1950, Watson and Signor 1986, Janssen 1991, Ohno et al. 1995, Carter and Schneider 1997, Persselin 1998, Schneider 1998), other aptations were based on field observations conducted during the course of this study and tested (e.g., surface area to volume). Photosymbiotic adaptations were de termined by ancestral trait reconstruction. Unless discussed otherwise, synapomorphic character states uniting all or some photosymbiotic fragines were considered adaptations for photosymbiosis, while symplesiomorphic character trait states were scored as exaptations (Gould and Vrba 1982). Windows The two species with confirmed window shell microstructure and phylogenetic representation ( C . cardissa and Fragum mundum ) were sister specie s, supporting a single origin of window shell micros tructure (Fig. 3-8). Howeve r, considering window shell microstructure in a broader sense (e.g., includ ing species with uncertain characters states, Table 3-3) alters this scenario. F . aff. mundum exhibits window-like shell transcluency in the posterior region, yet does not fall within windowed clade PSIII (Fig. 3-8) but instead is a well-supported member of PSII (the Â“ 25-ribÂ” group) (Fig. 38). If window shell microstructure is confirmed in this new sp ecies, it would lend support to a dual origin of window shell microstructure. Inclusion of windowed F. nivale in the phylogeny would further test the Â“single microstructura l windowÂ” hypothesis. Morphologically most similar to F. mundum , F. nivale is expected to fall in the PSIII subclade (Fig. 3-8) along with all other window ed species except F . aff. mundum . Documentation of window shell
78 microstructure in juvenile L. hemicardia , L. retusa and/or L. sp. 1 (Persselin 1998) from the PSIII subclade, coupled with the expected placement of F. nivale in this subclade, suggests that windows may have evolved in a common ancestor of PSI II (Fig. 3-9); all but F. unedo exhibit some degree of microstructu ral window formation. The evolution of pigment windows is similar to window shell mi crostructure, as all species that exhibit decreased pigmentation also possess microstructural windows (compare Figs. 3-8 & 3-9). The common ancestor of Clade III did not have pigment windows. An equivocal ancestral photosymbiotic state is rec overed for microstructural windows. Windows evolved after the origin of photosymbiosis but are cons idered an adaptation for photosymbiosis because they: 1) increase light pe netration to the shell interior and 2) are limited to a subset of photosymbiotic sp ecies (Watson and Signor 1986, Carter and Schneider 1997, Persselin 1998). Maximum observed adult size Small size was recovered as the most parsimonious ancestral character state for photosymbiotic fragines and all sampled fr agines (Fig. 3-10). Small size affords increased shell translucency and surface area to volume ratios and, because of these benefits, was hypothesized as aptive for a photosymbiotic lifestyle. Small size is considered an exaptation for photosymbi osis because it was recovered in nonphotosymbiotic species (e.g., symplesiomorphic) and clearly preceded the evolution of photosymbiosis.
79 Figure 3-8 Origin of photosymbiosis (dark bar) and window shell microstr ucture (WSM) traced onto a subset of the Fraginae phylogeny shown in Figure 3-7
80 Figure 3-9 Origin of photosymbiosis (dark bar) and decreased shell pigmentation or pi gment windows (PIGW) traced onto a subset of the Fraginae phylogeny shown in Figure 3-7.
81 Habitat Although the bulk of photosymbiotic species oc cupy a shallow, reef flat habitat, a deeper, non-reefal habitat was reconstructed as the most pa rsimonious ancestral character state for photosymbiotic fragines (as well as for all tested fragines, Fig. 3-11). Shallow reef flat habitats have the highest light intensity and th ese are considered aptive for photosymbiotic species. However, besides th e bulk of photosymbiotic fragines, two geographically disjunct, non-photosym biotic species also live in reef flat or reef flat-like environments, Microfragum festivum and A. media (Figs. 3-3 & 3-11). Because reef flat species were widely distributed phylogene tically (and geographi cally, given disjunct distributions of IWP M . festivum + photosymbiotic fragines, compared to west Atlantic A. media ), it is likely that a shallow reef flat ex istence evolved independe ntly in these three lineages/species. Exposure was reconstructed as shallowly infaunal for all nodes except that uniting: 1) photosymbiotic C. cardissa and F. mundum and 2) non-photosymbiotic P. vroomi and P. exiguum , the four epifaunal species (Fig. 3-12). Although an epifaunal existence is common to photosymbiotic and nonphotosymbiotic sister species, it has likely evolved independently in the two groups, given the large number of infaunal species sepa rating them. An epifaunal microhabitat is aptive because it allows increased light e xposure relative to an infaunal lifestyle. Although it has evolved independently in a non-photosymbiotic fr agine lineage, an epifaunal habitat is considered adaptive for C. cardissa and F. mundum because 1) it is also observed in the only othe r photosymbiotic cardiid lineage (Tridacninae) and 2) it is otherwise very rare in heterodont bivalves.
82 Figure 3-10 Origin of photosymbiosis (dar k bar) and maximum adult size (cm) traced onto a subset of the Fraginae phylogeny shown in Figure 3-7.
83 Figure 3-11 Origin of photosymbiosis (dar k bar) and habitat types (see Figure 3-3) traced onto a subset of the Fraginae phylogeny shown in Figure 3-7.
84 Shell measurements The most likely ancestral state of photosym biotic fragines was the same as that reconstructed for the subfamily as a whole: a modest (Category 4:100-120Â°, Fig. 3-13) keel. Flattening was difficult to reconstruc t due to the high variability exhibited, especially among Clade III members that ran the gamut of observed character state categories (Categories 1-6, Fig. 3-14). Alt hough the highest surface area to volume ratios were observed among photosymbiotic specie s, a relatively low surface area to volume ratio was supported as the most parsimonious ancestral character state among Clade III members, and also at deeper phylogenetic le vels (Category 2: 0.120.14) (Fig. 3-15). As with posterior shell flattening, photosymbiotic species exhibited the highest and lowest surface area to volume ratios, w ith the most elevated ratio s considered aptive for a photosymbiotic existence. All but one incide nce of elevated surf ace area to volume ratios (e.g., Categories 3-6) was specifi c to photosymbiotic fragines in the PSIII subclade (Fig. 3-15), suggesting that this state is adaptive for photosymbiosis. Increased taxon sampling permitted additiona l insight into surface area to volume trends among photosymbiotic fragines (Fig. 35). Photosymbiotic species that exhibited the lowest surface area to volume ratios, overlapping extensively with nonphotosymbiotic species, were: 1) restricted to the earliest diverging and 25-rib subclades (Fig. 3-15, PSI & PSII) and 2) from the sma llest size categories (Fig. 3-10, Categories 1&2).
85 Figure 3-12 Origin of photosymbiosis (dark bar) and microhabitat (exposure) traced onto a subset of the Fraginae phylogeny show n in Figure 3-7.
86 Figure 3-13 Origin of photosymbiosis (dark ba r) and mean keel angle (degrees) (see Figur e 3-2C) traced onto a subset of the Fra ginae phylogeny shown in Figure 3-7.
87 Figure 3-14 Origin of photosymbiosis (dark bar) and mean posterior she ll flattening (degrees)(see Fi gure 3-2B) traced onto a su bset of the Fraginae phylogeny shown in Figure 3-7
88 Figure 3-15 Origin of photosymbiosis (dark bar) and mean linearized surf ace area to volume ratios (see Figure 3-2D) traced onto a subset of the Fraginae phylogeny shown in Figure 3-7.
89 Discussion Photosymbiotic Status That all members of Fragum , Lunulicardia , and Corculum have algal symbionts, while all other fragine genera lack them, cont rasts with SchneiderÂ’s prediction (1998) that all Fraginae would be photosymbiotic (Table 3-2). Schneider (1998) based this proposal on observed digestive simplifications, however these are clearly not good indicators of photosymbiotic status, as the bulk of fragin es are not photosymbio tic. Therefore, these features need to be reconsidered (see belo w). The discovery of phot osymbiosis limited to three fragine genera confirms Persseli nÂ’s hypothesis (1998) of a paraphyletic and photosymbiotic Fragum giving rise to Lunulicardia and Corculum . Placing photosymbiotic status into a phylogenetic cont ext suggests a single a nd late acquisition of photosymbiosis in the Fraginae (Fig. 3-7). Character Trait Analyses and Evolution Windows The most interesting putative adaptati on for photosymbiosis is window shell microstructure, relatively transclucent shell regions restricted to the posterior shell surface. Conservative estimates support a single origin of microstructural windows uniting Fragum mundum and Corculum cardissa . However, windowed F. nivale , (unavailable for phylogenetic recons tructions), and putatively-windowed F . aff. mundum have the power to falsify this initial findi ng. The latter occurrence hints at multiple origins of window shell microstr ucture, or a deeper origin fo llowed by loss in intervening window-less species.Confirmation of window shell microstructure (and similarly decreased shell pigmentation) via rigorous shell analyses (Carter and Schneider 1997, Persselin 1998) in juveniles of L. hemicardia and L. retusa would not only enlarge the
90 windowed subclade, but also provide an intere sting twist to the evolution of window shell microstructure. Ontogenetic variation in microstructural and/or pigment windows suggested by preliminary observations in sampled L. hemicardia and L. retusa (and also Persselin 1998 for L. sp. 1) suggests that windows may be present at early life stages. Although not yet exhaustively samp led, adults of one of these ( L. retusa , Carter and Schneider 1997) did not exhibi t window microstructure, su ggesting that windows can be lost/overgrown with age. Based on the findings of this study, it is hypothesized that a more rigorous and comprehensive analysis of these prelimin ary observations of microstructural and pigment windows in juvenile and adult photosymbiotic fragines will support: 1) a deeper origin of shell modificati ons to enhance shell translucency that unites photosymbiotic fragines in PSIII (Fig. 3-8) and also, 2) a dual origin due to the presence of windows in F . aff. mundum , a well-supported member of the 25-rib clade (Fig. 3-8 PSII). Maximum observed adult size Ancestral reconstructions imply that the earliest photosymbiotic fragines were small (Category 1, 0.1-1 cm), suggesting a histor y of miniaturization in the origin of the group. The occurrence of distantly related nonphotosymbiotic fragines of small size (e.g., Microfragum , Parvicardium and Papillicardium , 0.1-1 cm) suggests this condition is symplesiomorphic and points to a history of miniaturi zation at deeper phylogenetic levels. The observed digestive simplifications documented among generic representatives by Schneider (1998), instead of being interp reted as indicative of a photosymbiotic lifestyle, are here reinterpreted simply as a consequence of small size (as observed in other groups, such as galeommatoids (Gir ibet and Wheeler 2002)). Moreover, this interpretation is consistent with a single origin of photosymbiosis uniting a relatively
91 small subclade of fragines (Clade III, Fig. 37), in contrast to earl ier suggestions of an entirely photosymbiotic subfamily (Schneid er 1998). Small size may have been an exaptation that facilitated the evolution of photosymbiosis: small animals are more shallowly infaunal than large ones, and also have thinner and thus more transclucent shells facilitating light penetration. Habitat The most likely ecological attr ibutes of an ancestral phot osymbiotic fragine include a shallowly infaunal, non-reef habitat (Fig s. 3-11 & 3-12). Only one photosymbiotic fragine, F. erugatum, occupies a non-reefal habitat, with all but one other species ( F. sueziense from turbid, lagoonal settings) living on th e reef flat (Fig. 3-3). This habitat is considered adaptive for a photosymbiotic lifes tyle, as is an epif aunal existence, a microhabitat occupied by just two photosymbiotic fragines, Corculum cardissa and F. mundum (Fig. 3-12). Shell measurements In contrast to earlier st udies that documented two ve ry divergent morphological ends in photosymbiotic fragines , the Â“tridacna-likeÂ” strategy of Fragum unedo juxtaposed against the Â“s olar panelÂ” strategy of Corculum cardissa (Ohno et al. 1995), this study provides evidence of a wide range of intermediate forms (Figs. 3-13 through 315). Among these intermediate fo rms, surface area to volume ratios were little elevated and many species interdigitated with a broa d sweep of non-photosymbiotic fragines and cardiids (Figs. 3-4C & 3-5). These trends were also revealed for k eel and posterior shell flattening measurements (that were found to be significantly correla ted with surface area to volume ratio) and photosymbiotic species also exhibited a wide range of size categories, overlapping with sampled nonphotosymbiotic species (Fig. 3-4A&B).
92 Principal component analyses among a broad suite of photosymbiotic and nonphotosymbiotic species revealed a Â“threshol dÂ” that separated a ll non-photosymbiotic from seven photosymbiotic species along pr incipal component 1 (along which species became flatter and more keeled, see dashed li ne on Fig. 3-6). This pattern suggests that, although many less flattened and less keeled non-photosymbiotic and photosymbiotic species co occur at below approximately 1, only photosymbiotic species moved beyond this threshold. Shell measurement trends highlight two im portant points regardi ng the evolution of form in photosymbiotic fragines. First, ra pid morphological evolution in response to photosymbiosis has occurred in fragines, where forms little different from nonphotosymbiotic cardiids (e.g., F. sueziense ) are closely related to extremely modified ones (e.g., C. cardissa ). This is best illustrated by ex amination of morphological patterns among photosymbiotic sister-species. For example, although both C. cardissa and F. mundum are windowed, the former is the second-lar gest sampled fragine, while the latter falls in the smallest size category. This sugge sts that, while the e volution of windows is conservative, size is highly labile. Second, because a wide range of photosymbio tic species is little differentiated in shell characters from non-photosymbiotic forms, these findings suggest that strategies for a photosymbiotic lifestyle besides shell modifications (e.g., as observed in C. cardissa ) are important. The relatively small size (0.11 cm) of certain phot osymbiotic species Â“naturallyÂ” affords elevated surface area to volume ratios, increased shell translucency and shallow infaunal burial (Figs.3-5, 310, 3-11 & 3-12). Anatomical and/or shell modifications that depart drastically from a basic bivalve bauplan are not necessary to
93 achieve an optimal light environment for zooxanthellar symbionts in small species. However, whereas small photosymbiotic clams are the least modified ( F. carinatum, F. sueziense and F. erugatum , with F. mundum an obvious exception), large photosymbiotic clams have the most drastically al tered and pronounced morphologies ( F. unedo, C. cardissa, L. hemicardia and L. retusa ). This suggests that be ing large and photosymbiotic necessitates major modifications to ecology, behavior and/or anatomy (both shell and tissues) in order to provide adequate light levels to photosymbionts, similar to trends observed in giant clams (Yonge 1936).
94 CHAPTER 4 GLOBAL DISCUSSION Molecular phylogenetics These results support the original Fragin ae circumscribed by Stewart (1930) and are in almost complete agreement with th e subfamilial delineation of Keen (1980) and Voskuil and Onverwagt (1989). Circumscripti ons differ in that both Keen (1980) and Voskuil and Onverwagt ( 1989) included the genus Afrocardium in the Fraginae (excluded from the subfamily in these analyses) and Ctenocardia victor is best supported as a distantly related cardiid by this study and not a fragine. The present results contrast significantly with the Fraginae as circumscribed by Schneider (1998), largely because of the high levels of sequence divergence that resulted in very different placement of Parvicardium sensu lato representatives. These an alyses do not support the inclusion of divergent P. minimum and Papillicardium papillosum (Clade IV) in the Fraginae. That Clade IV is distantly related to all tested fragines is suppor ted by the recovery of one of the tested outgroup clades as more closely re lated to fragine lineag es (Clades I-III) than P. minimum + Papillicardium papillosum . However, although it is a pparent that Clade IV is distantly related to the originally circum scribed subfamily (Clades II and III), whether the remainder of sampled Parvicardium (Clade I) is supported as members of a divergent Fraginae is equivocal. The inclus ion of additional species of Parvicardium but, more importantly, appropriate molecular markers for resolution at deeper phy logenetic levels is necessary before circumscribing a new Fraginae.
95 The lack of a perisiphonal suture results in confluence of the incurrent siphonal aperture and pedal gape. This strongest of the morphological synapomorphies uniting the Fraginae may be related to miniaturi zation, not a photosymbi otic lifestyle. Schneider (1998) recognized the distinctiveness of Papillicardium and recovered a paraphyletic Parvicardium . The findings presented here support Papillicardium as distinct from all other shallow-water Mediterranean Parvicardium sampled in this study (Clade I). Although Papillicardium papillosum is recovered as sister to Parvicardium minimum, this may be due to long-branch attrac tion; both taxa were highly divergent for all gene regions sequenced. P. minimum is a deep-water species and the sample included from this study was from Sweden, ge ographically disjunct from other small Parvicardium and P. papillosum . Clade I is composed of three recognized species and one highly differentiated taxon (based on a single individual from La Herra, Spain) from the genus Parvicardium . As P. exiguum, considered the type species in the most recent and comprehensive review of the Mediterranean Parvicardium (van Aartsen and Goud 2000, but also Hylleberg 2004 and others), is a member of this clade, the nomen Parvicardium should be restricted to this group. These preliminary results suggest that sm all European cardiids, presently allied to Parvicardium sensu lato, are highly divergent li neages, as recognized by others (e.g. Schneider (1998) supporte d generic status of Papillicardium papillosum ). Given these findings, geographically disjunc t but presently unsampled Parvicardium species ( P. turtoni from South Africa and P. pinnulatum from the west Atlantic ) are not expected to fall within this clade that is currently restricted to morphologically similar species from relatively shallow Mediterranean habitats. The equivocal findings generated by the
96 phylogenetic results of this study call atte ntion to the need for a comprehensive systematic revision of members currently allied to the earliest dive rging fragine genera. The genera Trigoniocardia and Ctenocardia were recovered as a well-supported clade in Bayesian analyses of both mitochondrial and combin ed datasets. This contrasts with Schneider (1998), who split these genera into two divergent fragine subclades and recovered the Trigoniocardia group as sister to all fragines (other than Parvicardium sensu lato). The genus Americardia is most morphologically similar to Ctenocardia not Trigoniocardia, as recognized recently by Schneider (1998) as well as Voskuil and Onverwagt (1989), who synonymized Americardia with Ctenocardia because of phenetic similarity. However, in contrast with Schneider (1998) and Voskuil and Onverwagt (1989), but consistent with the present analys es, is the bulk of earlier taxonomic work (Stewart 1930, Clench and Smith 1944, Keen 1951, 1980, Olsson 1961 and Popov 1977) that recognized Americardia as a subgenus of Trigoniocardia . The present molecular phylogenetic findings, along with these older works, suggest that Clade II may be composed of separate IWP (Ctenocardia-Microfragum) and American (AmericardiaTrigoniocardia) radiations. The morphological disparity of C. victor, relative to other Ctenocardia, has long been appreciated (Wilson a nd Stevenson). The high levels of sequence divergence in all four gene regions reported here for multiple individuals, coupled with the morphological disparity obser ved by others, confirm the distinctiveness of this species. Clade III is composed of members of the genera Fragum, Lunulicardia and Corculum . In all analyses, Fragum is strongly supported as para phyletic and in combined and mitochondrial datasets is composed of three subclades. The taxonomic affinities of F.
97 erugatum and F. sueziense have been long debated, with both species allied to no fewer than five different genera (see Hylleberg 2004). The results of this study unite these two as sisters, sister to all other tested Fragum , Lunulicardia and Corculum species. Like all other tested fragine species, F . aff . mundum from Tuvalu is reciprocally monophyletic. Initially thought to be a variant of F. mundum following preliminary morphological examination, F. aff. mundum was recovered as a member of the 25-rib clade and sister to F. carinatum . Preliminary findings support Corculum cardissa as a single species, with morphological variability better explained in the context of ontogeny or local adaptation to environmental conditions (Kawaguti 1950). Geographic structuring was evident among all tested Fragum species targeted for intensive population-level sampling. For ex ample, individuals from Okinawa of F. loochoanum, F. unedo and F. mundum were consistently r ecovered as well-supported, reciprocally monophyletic linea ges compared with populations sampled from all other locales. Preliminary evidence of fine-scale geographic structuring, coupled with poor dispersal capabilities (e.g. for tested Corculum from Palau, Kawaguti 1950), suggests that this group is a good bivalve candidate for futu re phylogeographic initiat ives in the insular IWP (Paulay and Meyer 2002, Kirkenda le and Meyer 2004, Meyer et al. 2005. Character trait evolution That all members of Fragum, Lunulicardia, and Corculum have algal symbionts, while all other fragine genera lack them, cont rasts with SchneiderÂ’s prediction (1998) that all Fraginae would be photosymbiotic. Schneid er (1998) based this proposal on observed digestive simplifications, however thes e are clearly not good indicators of photosymbiotic status, as the bulk of fr agines are not photosymbiotic. Placing
98 photosymbiotic status into a phylogenetic cont ext suggests a single a nd late acquisition of photosymbiosis in the Fraginae. The most interesting putative adaptati on for photosymbiosis is window shell microstructure, relatively transclucent shell regions restricted to the posterior shell surface. Conservative estimates support a single origin of microstructural windows uniting Fragum mundum and Corculum cardissa. Ontogenetic variation in microstructural and/or pigment windows s uggested by preliminary observations in sampled L. hemicardia and L. retusa (and also Persselin 1998 for L . sp. 1) suggests that windows may be present at early life stages . Based on the findings of this study, it is hypothesized that a more rigorous and compre hensive analysis of these preliminary observations of microstruc tural and pigment windows in juvenile and adult photosymbiotic fragines will support: 1) a deeper origin of shell modifications to enhance shell translucency that unites photosymbiotic fragines in PSIII and also, 2) a dual origin due to the presence of windows in F. aff. mundum , a well-supported member of the 25rib clade (PSII). Ancestral reconstructions imply that the earliest photosymbiotic fragines were small, suggesting a history of miniaturizati on in the origin of the group. However, the occurrence of distantly related non-photosymbio tic fragines of small size suggests this condition is symplesiomorphic and points to a history of miniaturization at deeper phylogenetic levels. The observed digestive simplifications documented among generic representatives by Schneider ( 1998) are here reinterpreted s imply as a consequence of small size. Moreover, this interpretation is consistent with a single origin of photosymbiosis uniting a relatively small subcla de of fragines, in contrast to earlier
99 suggestions of an entirely photosymbiotic subfamily (Schneider 1998). Small size may have been an exaptation that facilitated th e evolution of photosymbiosis: small animals are more shallowly infaunal than large one s, and also have thinner and thus more transclucent shells facili tating light penetration. The most likely ecological attr ibutes of an ancestral phot osymbiotic fragine include a shallowly infaunal, non-reef habitat This habitat is considered adaptive for a photosymbiotic lifestyle, as is an epifaunal existence, a micr ohabitat occupied by just two photosymbiotic fragines, Corculum cardissa and F. mundum . In contrast to earlier st udies that documented two ve ry divergent morphological ends in photosymbiotic fragines , the Â“tridacna-likeÂ” strategy of Fragum unedo juxtaposed against the Â“s olar panelÂ” strategy of Corculum cardissa (Ohno et al. 1995), this study provides evidence of a wide range of intermediate forms. Among these intermediate forms, surface area to volume ratios were little elevated and many species interdigitated with a broad sweep of non-phot osymbiotic fragines and cardiids. These trends were also revealed for keel and posteri or shell flattening measurements (that were found to be significantly correla ted with surface area to vol ume ratio) and photosymbiotic species also exhibited a wi de range of size categories, overlapping with sampled nonphotosymbiotic species. Principal compone nt analyses among a broad suite of photosymbiotic and non-photosymbiotic species revealed a Â“thresholdÂ” that separates all non-photosymbiotic from seven photosymbiotic species. This pattern suggests that, although many less flattened and less keeled non-photosymbiotic and photosymbiotic species co occur at below approximately 1, only photosymbiotic species moved beyond this threshold.
100 Shell measurement trends highlight two im portant points regardi ng the evolution of form in photosymbiotic fragines. First, ra pid morphological evolution in response to photosymbiosis has occurred in fragines, where forms little different from nonphotosymbiotic cardiids (e.g. F. sueziense ) are closely related to extremely modified ones (e.g. C. cardissa ). This is best illustrated by ex amination of morphological patterns among photosymbiotic sister-species. For example, although both C. cardissa and F. mundum are windowed, the former is the second-lar gest sampled fragine, while the latter falls in the smallest size category. This sugge sts that, while the e volution of windows is conservative, size is highly labile. Second, because a wide range of photosymbio tic species is little differentiated in shell characters from non-photosymbiotic forms, these findings suggest that strategies for a photosymbiotic lifestyle besides shell modifications (e.g. as observed in C. cardissa ) are important. The relatively small size of certain photosymbiotic species Â“naturallyÂ” affords elevated surface area to volume ratios, increased shell translucency and shallow infaunal burial. Anatomical and/or shell modifi cations that depart drastically from a basic bivalve bauplan are not necessary to ach ieve an optimal light environment for zooxanthellar symbionts in small species. Ho wever, whereas small photosymbiotic clams are the least modified ( F. carinatum, F. sueziense and F. erugatum , with F. mundum an obvious exception), large photosymbiotic clam s have the most drastically altered and pronounced morphologies ( F. unedo, C. cardissa, L. hemicardia and L. retusa ). This suggests that being large a nd photosymbiotic necessitates major modifications to ecology, behavior and/or anatomy in order to provide adequate light levels to photosymbionts, similar to trends observed in giant clams (Yonge 1936).
APPENDIX A NUMBER OF INDIVIDUALS EXAMINED FOR SCORED CHARACTERS OF FRAGINAE
102No Species WSM PIGW SIZE HAB EXP DEPTH KEEL FLAT SAV 1 Fragum fragum P1998 P1998 10 10 10 10 3 3 5 2 Fragum scruposum 5 5 10 10 10 10 4 4 5 3 Fragum loochoanum P1998 P1998 10 10 10 10 3 3 5 4 Fragum carinatum 5 5 10 10 10 10 3 3 5 5 Fragum aff mundum 2 2 4 4 4 4 2 2 2 6 Fragum mundum P1998 P1998 10 10 10 10 1 1 5 7 Fragum unedo P1998 P1998 10 10 10 10 4 4 7 8 Fragum erugatum 5 5 10 10 10 10 3 3 5 9 Fragum sueziense P1998 P1998 10 10 10 10 2 2 5 10 Lunulicardia retusa CS1997 CS1997 10 10 10 10 3 3 5 11 Lunulicardia hemicardia 5 5 10 10 10 10 3 3 5 12 Corculum cardissa CS1997 CS1997 10 10 10 10 4 4 8 13 Americardia media P1998 P1998 10 10 10 10 3 3 5 14 Americardia biangulata 1 1 2 2 2 2 1 1 1 15 Trigoniocardia granifera CS1997 CS1997 10 10 10 10 3 3 5 16 Apiocardia obovale 2 2 3 3 3 3 2 2 2 17 Microfragum festivum P1998 P1998 10 10 10 10 3 3 5 18 Ctenocardia fornicata P1998 P1998 10 10 10 10 3 3 4 19 Ctenocardia gustavi 2 2 5 5 5 5 2 2 2 20 "Ctenocardia" victor 4 4 10 10 10 10 3 3 4 21 Papillicardium papillosum 5 5 10 10 10 10 3 3 5 22 Parvicardium vroomi 5 5 10 10 10 10 3 3 5 23 Parvicardium exiguum 5 5 10 10 10 10 3 3 5 P1998 refers to Persselin 1998, CS1997 re fers to Carter and Schneider 1997
APPENDIX B PHOTOSYMBIOTIC STATUS, MORPHOMETRIC AND HABITAT DATA FOR MEASURED CHARACTERS
104 No Species size (cm)groupsrange (m)1pref. depth (m) groups Mean Flat groups Mean Keel groupsMean SAVGroups 1 Fragum fragum 4.2 5 0-10 1 1 113.836 88.17 3 0.149 2 2 Fragum scruposum 1.9 2 0-3 1 1 98.2 3 99.9 3 0.105 1 3 Fragum loochoanum 1.7 2 0-2 0.5 1 93.83 2 104 4 0.122 2 4 Fragum carinatum 0.5 1 0-1 0.3 1 107.3 5 102.83 4 0.131 2 5 Fragum aff mundum 0.8 1 0-2 2 1 124.5 7 97.5 3 0.135 2 6 Fragum mundum 0.6 1 0-1 0.5 1 126 7 86 3 0.136 2 7 Fragum unedo 7.9 8 0-5 0.5 1 114.636 105.25 4 0.147 2 8 Fragum erugatum 1.9 2 0-2 0.5 1 87.3 1 116.83 4 0.114 1 9 Fragum sueziense 0.6 1 15-20 18 2 99 3 123.25 5 0.125 2 10 L unulicardia retusa 4.2 5 0-10 3 1,2 113.3 6 96.17 3 0.171 3 11 L unulicardia hemicardia 4.1 5 0-2 0.5 1 125.3 7 86.83 3 0.16 3 12 Corculum cardissa 5.2 6 0-2 0.5 1 135 7 40.25 1 0.22 5 13 A mericardia media 3.6 4 ? 1 2,3 105 4 113.67 4 0.124 2 14 A mericardia biangulata 1.7 2 ? 30 2,3 96 3 92 3 0.124 2 15 Trigoniocardia g ranifera 1.2 2 0-30 1 1,2 87.3 1 96.7 3 0.126 2 16 A piocardia obovale 1.4 2 0-30 1 2,3 101.5 4 83.5 3 0.135 2 17 Microfragum festivum 1 1 0-5 5 1,2 100.334 126.33 5 0.119 1 18 Ctenocardia fornicata 3.6 4 ? 25 2,3 105.5 5 116 4 0.151 3 19 Ctenocardia gustavi 2.1 3 20-40 50 2,3 114.5 6 107.5 4 0.123 2 20 Ctenocardia victor 2.5 3 30-100 100 3 93 2 119 4 0.131 2 21 P apillicardium p apillosum 1 1 3-10 15 2 4 5 1 22 P arvicardium vroomi 0.7 1 0-3 2 1 108.675 106 4 0.119 1 23 P arvicardium exiguum 1.1 2 ? 1 1 113.3 6 102.83 4 0.12 1 1 Taken from the literature, personal observation and museum collections
APPENDIX C RAW LOGGED DATA FOR 95% CONFIDENCE INTERVAL PLOTS OF KEEL ANGLE, POSTERIOR SHELL FLATTENING AND LINEARIZED SURFACE AREA TO VOLUME RATIOS
106Species FraginaeP/Smean keelupper keellower keelmean flat upper flatlower flatmean savupper savlower sav Corculum cardissa Yes Yes3.68392 3.95249 3.41535 4.9041 4.9934 4.81481-1.515 -1.55971-1.47029 Fragum erugatum Yes Yes4.75783 4.99243 4.52323 4.46861 4.61277 4.32446-2.17532-2.25439-2.09624 Lunulicardia retusa Yes Yes4.56343 4.78699 4.33987 4.72904 4.88344 4.57464-1.76665-1.8514 -1.6819 Lunulicardia hemicardia Yes Yes4.46366 4.54177 4.38555 4.83 4.96404 4.69597-1.83357-1.95215-1.71499 Fragum fragum Yes Yes4.47918 4.50858 4.44978 4.73378 4.86755 4.60001-1.90791-2.04461-1.77122 Fragum unedo Yes Yes4.65075 4.84159 4.45992 4.73913 4.86999 4.60827-1.92262-1.99773-1.84752 Fragum loochoanum Yes Yes4.64211 4.84607 4.43815 4.54109 4.63015 4.45204-2.11628-2.34239-1.89017 Fragum scruposum Yes Yes4.60354 4.65293 4.55414 4.58621 4.64173 4.53068-2.25435-2.28131-2.22739 Fragum carinatum Yes Yes4.63211 4.76874 4.49548 4.67576 4.73388 4.61763-2.02909-2.05066-2.00753 Trigoniocardia granifera Yes No4.57069 4.67409 4.46729 4.46878 4.60229 4.33527-2.07066-2.17405-1.96727 Ctenocardia victor Yes No4.7386 4.83642 4.64078 4.53154 4.67108 4.392 -2.03315-2.14208-1.92422 Microfragum festivum Yes No4.83769 4.98949 4.68589 4.60832 4.66513 4.55151-2.1347 -2.33623-1.93317 Ctenocardia fornicata Yes No4.75264 4.88496 4.62032 4.65823 4.75258 4.56389-1.89349-1.96339-1.82359 Papillocardium papillosum Yes No4.82919 5.00994 4.64844 4.67564 4.74949 4.6018 -2.12278-2.17501-2.07056 Parvicardium vroomi Yes No4.66259 4.78822 4.53697 4.68779 4.7838 4.59178-2.13262-2.19156-2.07368 Americardia media Yes No4.7327 4.83505 4.63035 4.70195 4.73029 4.67361-2.0841 -2.14578-2.02242 Parvicardium exiguum Yes No4.63274 4.71561 4.54987 4.72879 4.89824 4.55935-2.12302-2.17199-2.07405 Cerastoderma edule No No4.75261 4.88804 4.61718 4.46929 4.5599 4.37868-2.3121 -2.37755-2.24665 Vasticardium orbita No No4.90763 4.95374 4.86151 4.49005 4.58187 4.39823-2.06192-2.24973-1.87411 Acanthocardia echinata No No4.78709 4.83879 4.7354 4.53152 4.61647 4.44658-2.35392-2.43423-2.27362
107 LIST OF REFERENCES Bartsch P. 1947. The little hearts ( Corculum ) of the Pacific and Indian oceans. Pacific Science 1 : 221-228. Beckage NE. 1998. Parasitoids and polydnaviruses. BioScience 48(4) : 305-311. Bernard FR. 1972. The genus Thyasira in Western Canada (Bivalvia: Lucinacea). Malacologia 11 : 365-389. Berner T, Wishkovsky A, Dubinksy Z. 1986. Endozoic algae in shelled gastropods-A new symbiotic association in coral reefs? II. Survey of distribution of endozoic algae in Red Sea snails. Coral Reefs 5(2): 107-109. Bouchet P, von Cosel R. 2004. The world's largest lucini d is an undescribed species from Taiwan (Mollusca: Bivalvia). Zoological Studies 43(4) : 704-711. Bourtzis K, OÂ’Neill S. 1998. Wolbachia infection and ar thropod reproduction. BioScience 48(4) : 287-293. Buddemeier RW, Fautin DG. 1993. Coral bleaching as an adaptive mechanism-A testable hypothesis. BioScience 43(5) : 320-326. Carter JG, Schneider JA. 1997. Condensing lenses and shell microstructure in Corculum (Mollusca: Bivalvia). Journal of Paleontology 71(1) : 56-61. Chomczynski P, Mackey K, Drews R, Wilfinger W. 1997. DNAzol: A reagent for the rapid isolation of genomic DNA. BioTechniques 22 : 550-553. Clench WJ, Smith LC. 1944 . The family Cardiidae in the western Atlantic. Johnsonia 13 : 32pp. Coan EV, Scott PV, Bernard FR. 2000 . Bivalve Seashells of Western North America. Marine bivalve molluscs from Arctic Alaska to Baja California. Santa Barbara Museum of Natural History Monogr aphs 2, Studies in Biodiversity 2: 1-764. Collin R. 2004. Phylogenetic effects, the loss of comp lex characters, and the evolution of development in calyptraeid gastropods. Evolution 58(7): 1488-1502. Corliss JB, Cymond J, Gordon LI, Edmond JM, von Herzen RP, Ballard RD, Green K, Williams D, Bainbridge A, Carne K, van Andel TH. 1979. Submarine thermal springs on the GalÃ¡pagos Rift. Science 203: 1073-1083.
108 Cowen R. 1983. Algal symbiosis and its recogni tion in the fossil record. In Biotic Interactions in Recent and Fossil Benthic Communities (MJS Tevesz & PL McCall, eds.). New York, NY: Plenum Press. Dana JD. 1846. Zoophytes . United States Expl oring Expedition 1838-1842: 7 . Distel DL. 1998. Evolution of chemoautotrophic endosymbioses in bivalves. BioScience 48(4): 277-286. Douglas AE. 1994. Symbiotic Interactions . Oxford: Oxford University Press. Dufour SC, Felbeck H. 2003. Sulphide mining by the supere xtensile foot of symbiotic thyasirid bivalves. Nature 426: 65-67. Farmer MA, Fitt WK, Trench RK. 2001. Morphology of the symbiosis between Corculum cardissa (Mollusca: Bivalvia) and Symbiodinium corculorum (Dinophyceae). Biological Bulletin 200 : 336-343. Felbeck H, Childress JJ, Somero GN. 1983. Biochemical interactions between molluscs and their algal and bacterial symbionts. In: The Mollusca, Volume 2 Environmental Biochemistry and Physiology (Hockhackha PW, ed.). New York, NY: Academic Press. Fitt WK, Trench RK. 1981. Spawning, development and acquisition of zooxanthellae by Tridacna squamosa (Mollusca: Bivalvia). Biological Bulletin 161: 213-235. Fitt WK, Fisher CR, Trench RK. 1986. Contribution of the symbiotic dinoflagellate Symbiodinium microadriaticum to the nutrition, growth a nd survival of larval and juvenile tridacnid clams. Aquaculture 55 : 5-22. Fitt WK, Helsinga GA, Watson TC. 1993. Utilization of dissolved inorganic nutrients in growth and maricultur e of the tridacnid clam Tridacna derasa . Aquaculture 109 (1): 27-38. Folmer O, Black M, Hoen W, Lutz R, Vrijenhoek R. 1994. DNA primers for amplification of mitochondr ial cytochrome c oxidase subunit I from diverse metozoan invertebrates. Molecular Marine Biology and Biotechnology 3: 294299. Fukami H, Budd AF, Paulay G, Sole-Cava A, Chen CA, Iwao K, Knowlton N. 2004. Conventional taxonomy obscures deep dive rgence between Paci fic and Atlantic corals . Nature 427 : 832-835. Futuyma D. 1998. Evolutionary Biology 3rd ed. Sunderland, MA: Sinauer Associates, Inc.
109 Giribet G, Wheeler W. 2002. On bivalve phylogeny: A high-level analysis of the Bivalvia (Mollusca) based on combin ed morphology and the DNA sequence data. Invertebrate Biology 121(4) : 271-324. Giribet G, Distel DL. 2003 . Bivalve phylogeny and molecular data. In: Molecular Systematics and Phylogeography of Mollusks (Lydeard C & Lindberg DR, eds.). Washington, DC: Smithsonian Books. Givnish T. 1997. Adaptive radition and molecular systematics: Issues and Approaches. In: Molecular Evolution and Adaptive Radiation (Givnish TJ & Systma KJ, eds.). Cambridge, MA: Cambridge University Press. Glover EA, Taylor JD. 1997. New species and records of Rastafaria and Megaxinus (Bivalvia : Lusicindae) from the Wester n Indian ocean and Red Sea, with a reappraisal of Megaxinus . Journal of Conchology 36 : 1-18. Glover EA, Taylor JD. 2001. Systematic revision of Australian and Indo-Pacific Lucinidae (Mollusca : Bivalvia): Pillucina , Wallucina and descriptions of two new genera and four new species. Records of the Australian Museum 53 : 263-292. Goreau TF, Goreau NI, Yonge CM. 1973. On the utilization of photosynthetic prodcuts from zooxanthellae and of a dissolved amino acid in Tridacna maxima f. elongata (Mollusca : Bivalvia). Journal of the Zoologi cal Society of London 169 : 417-454. Gould SJ, Lewontin RC. 1978. The Spandrels of San Marco and the Panglossian paradigm: A critique of th e adaptationist programme. Proceedings of the Royal Society of London 205: 581-598. Gould SJ, Vrba ES. 1982. ExaptationA missing term in the science of form. Paleobiology 8(1): 4-15. Gray MW, Burger G, Lang BF. 1999. Mitochondrial evolution. Science 283 : 14761481. Hartman MC, Pratt I. 1976. Infection of the heart cockle, Clinocardium nuttalli from Yaquina Bay, Oregon with an endosymbiotic alga. Journal of Invertebrate Pathology 28 : 291-299. Hickman CS. 1994. The genus Parvilucina in the Eastern Pacific: Making evolutionary sense of a chemosymbiotic species complex. Veliger 37 : 43-61. Hylleberg J. 2004. Lexical approach to Cardiacea. Phuket Marine Biological Center Special Publication 1-3 . Janssen HH. 1991 . The peculiar morphology and ultras tructure of the heart shell, Corculum cardissa (Bivalvia: Cardiacea: Fraginae), a consequence of adaptation to endosymbiotic zooxanthellae. Proceedings of the Tenth International Malacological Congress (Tubingen 1989) : 85-88.
110 Jones DJ, Jacobs DK. 1992. Photosymbiosis in Clinocardium nuttalli : Implications for tests of photosymbiosis in fossil molluscs. Palaios 7 : 86-95. Jumpponen A, Trappe JM. 1998. Dark-septate root endophyt es: a review with special reference to facultative biotrophic symbiosis. New Phytologist 140: 295-310. Kafanov AI, Popov SV. 1977. K sisteme kaynozoyiskikh kardiodey (Bivalvia). Paleontologicheskiy Zhurnal . 55-64. Kawaguti S. 1941. Heart shell Corculum cardissa (L.) and its zooxanthella. Kagaku Nanyo 3: 45-46 [In Japanese]. Kawaguti S. 1950. Observations on the heart shell, Corculum cardissa (L.), and its associated zooxanthellae. Pacific Science 4 : 43-49. Kawaguti S. 1968. Electron microscopy on zooxanthellae in the mantle and gill of the heart shell. Biological Journal of Okayama University 14(1-2): 1-11. Kawaguti S. 1983. The third record of associat ion between bivalve molluscs and zooxanthellae. Proceedings of Japan Academic Series B 59(2) : 17-20. Keen AM. 1951. Outline of a proposed classification of the pelecypod family Cardiidae. Minutes of the Conchological Club of Southern California 111 :6-8. Keen AM. 1980. The pelecypod family Cardiidae: A taxonomic summary. Tulane Studies in Geology and Paleontology 16(1) : 1-40. Kilburn R, Rippey E. 1982. Sea shells of southern Africa . MacMillan South Africa Ltd. Kirkendale LA. Character trait evolution in the marine clam subfamily Fraginae. Chapter 1. Kirkendale LA, Meyer CP. 2004 . Molecular phylogenetics of the Patelloida profunda group: Diversification in a dispersal-driven marine system. Molecular Ecology 13 (9) : 2749-2762. Kocher TD, Thomas WK, Meyer A, Edwards SV, Paabo S, Villablanca FX, Wilson AC. 1989. Dynamics of mitochondrial DNA evol ution in animals: Amplification and sequencing with conserved primers. Proceedings of the National Academy of Sciences USA 86 : 6196-6200. Koyama Y, Yamamoto T, Toki Y, Minato H. 1981. A catalogue of molluscs of Wakayama Prefecture, the Province of Kii I. Bivalvia, Scaphopoda and Cephalopoda. Special Publications from the Se to Marine Biological Laboratory, Special Publications Series VII . Lamprell K, Whitehead T. 1992. Bivalves of Australia 1 . Bathurst: Crawford House Press.
111 Lee JJ, Anderson OR. 1991. Symbiosis in Foraminifera. In: Biology of Foraminifera (eds. Lee JJ, Anderson OR). London: Academic Press. Lindquist, N, Barber PH, Weisz JB. 2005. Episymbiotic microbes as food and defence for marine isopods: Unique symbio ses in a hostile environment. Proceedings of the Royal Society BBiological Sciences 272 (1569): 1209-1216. Maddison WP, Maddison DR. 2000. MacClade.v. 4 . Sunderland, MA: Sinauer and Associates, Inc. Margulis L. 1998. Symbiotic Planet: A New Look at Evolution . New York, NY: Basic Books. Margulis L, Fester R. 1991. Symbiosis as a Source of Evolutionary Innovation: Speciation and Morphogenesis . Boston, MA: MIT Press. Maruyama T, Ishikura M, Ya mazaki S, Kanai S. 1998. Molecular phylogeny of zooxanthellate bivalves. Biological Bulletin 195 :70-77. Mayr E. 1963. Animal Species and Evolution . Cambridge, MA: Belknap Press of Harvard University Press. McFall-Ngai, MJ, Ruby, EG. 1998. Sepioloid and vibrios: When they first meet. BioScience 48(4) : 257-265. Meeks JC. 1998. Symbiosis between nitrogen-fix ing cyanobacteria and plants. BioScience 48(4) : 266-276. Meyer CP. 2003 . Molecular systematics of cowries (Gastropoda: Cypraeidae) and diversification patterns in the tropics. Biological Journal of the Linnean Society 79 : 401-459. Meyer CP. 2004. Toward comprehensiveness: Increased molecular sampling within Cypraeidae and its phyl ogenetic implications. Malacologia 46(1): 127-156. Meyer CP, Geller JB, Paulay G. 2005. Fine scale endemism on coral reefs: archipelagic differentiation in turbinid gastropods. Evolution 59 : 113-125. Moore RC, Teichert C, McCormick L, Williams RB. 1969. Treatise on Invertebrate Paleontology Part N, Vol. 2, Mollusca 6, Bivalvia. The Geological Society of America, Inc. and The University of Kansas. Moran NA, Telang A. 1998. Bateriocyte-associated symbionts of insects. BioScience 48(4): 295-304. Moritz C. 1994. Defining Â“Evolutionary Significant UnitsÂ” for conservation. Trends in Ecology and Evolution 9: 373-375.
112 Morton B. 1978. The diurnal rhythm and the proce sses of feeding and digestion in Tridacna crocea (Bivalvia: Tridacnidae). Journal of the Zoological Society of London 185: 371-387. Morton B. 1982. The biology, functional morphol ogy and taxonomic status of Fluviolanatus subtorta (Bivalvia: Trapeziidae), a heteromyarian bivalve possessing Â“zooxanthellaeÂ”. Journal of the Malacologi cal Society of Australia 5(3-4): 113-140. Morton B. 2000. The biology and functional morphology of Fragum erugatum (Bivalvia: Cardiidae) from Shark Bay, We stern Australia: The significance of its relationship with entrained zooxanthelllae. Journal of Zoology, London 251 :3952. Mujer CV, Andrews DL, Manhart JR, Pierce SK, Rumpho ME. 1996. Chloroplast genes are expressed during intracel lular symbiotic association of Vaucheria littorea plastids with the sea slug Elysia chlorotica . Proceedings of the National Academy of Sciences 93: 12333-12338. Muscatine L. 1967. Glycerol excretion by symbio tic algae from corals and Tridacna and its control by the host . Science 156 : 516-519. Naidu KS, South GR. 1970. Occurrence of an endozoic al ga in the giant scallop Placopecten magellanicus (Gmelin). Canadian Journal of Zoology 48: 183-185. Norris RD. 1996. Symbiosis as an evolutionary i nnovation in the radiation of Paleocene planktic foraminifera. Paleobiology 22: 461-480. Norton JH, Shepherd MA, Long HM, Fitt WK. 1992. The zooxanthellae tubular system in the giant clams. Biological Bulletin 183 : 503-506. Ohno T, Katoh T, Yamasu T. 1995 . The origin of algal-bivalve photosymbiosis. Palaeontology 38(1): 1-21. Oliver PG. 1992. Bivalved Seashells of the Red Sea . Verlag Christa Hemmen, National Museum of Wales. Olsson AA. 1961. Molluscs of the Tropical Eastern Pacific. Ithaca, NY: Paleontological Research Institution. Palumbi SR. 1996. Nucleic Acids II: The polymerase chain reaction. In: Molecular Systematics, 2nd edn. (Hillis DM, Moritz C & Mable BK, eds.) Sunderland, MA: Sinauer Associations, Inc. Pardy RL. 1980. Symbiotic algae and 14C incorporation in the freshwater clam Anodonta . Biological Bulletin 158: 349-355.
113 Park J-K, " Foighil D. 2000. Sphaeriid and corbiculid clams represent separate heterodont bivalve radiations into freshwater environments. Molecular Phylogenetics and Evolution 14(1): 75-88. Paulay G. 1997. Diversity and distribution of reef organisms. In: Life and death of coral reefs (Birkeland C, ed.). New York, NY: Chapman & Hall. Paulay G, Meyer C. 2002. Diversification in the tropi cal Pacific: Comparison between marine and terrestrial systems and the importance of founder speciation. Integrative & Comparative Biology 42 : 922-934. Payne CM, Allen JA. 1991. The morphology of deep-sea Thyasiridae (Mollusca: Bivalvia) from the Atlantic Ocean. Philosophical Transactions of the Royal Society, Series B 334 : 481-566. Persselin S. 1998. The Evolution of Shell Windows w ithin the Fraginae (Bivalvia: Cardiidae) and the Origin of Algal Symbiosis in Cardiids. Mangilao, Guam: University of Guam Marine Laboratory. Popov SV. 1977. Mikrostruktura rakovniy I sistema tika kardiid. Akademiya Nauk SSSR. Trudy Paleontogicheskigo Instituta 153 :1-124. Posada D, Crandall KA. 1998. MODELTEST: Testing the model of DNA substitution. Bioinformatics 14(9): 817-818. Poutiers JM. 1992. The Australasian Protocardiinae revisited (Bivalvia: Cardiidae). American Malacological Bulletin 9(2): 139-144. Reid RGB. 1990. Evolution implications of sulphi de-oxidizing symbioses in bivalves. In: The Bivalvia-Proceedings of a Memorial Symposium in Honor of Sir Charles Maurice Yonge , Edinburgh, 1986 (ed. Morton B.) Hong Kong: Hong Kong University Press. Reid RGB, Brand DG. 1986. Sulfide-oxidising symbiosis in lucinaceans: implications for bivalve evolution. Veliger 29: 3-24. Richardson SL. 2001. Endosymbiont change as a ke y innovation in the adaptive radiation of Soritida (Foraminifera). Paleobiology 27(2) : 262-289. Ronquist F, Huelsenbeck JP. 2003. MrBayes 3: Bayesian phylogenetic inference under mixed models. Bioinformatics 19 (12): 1572-1574. Rosewater J. 1965. The family Tridacnidae in the Indo-Pacific. Indo-Pacific Mollusca 1 (6 ): 347-396. Rowan R. 1998. Diversity and ecology of zooxanthellae on coral reefs. Journal of Phycology 34(3): 407-417.
114 Rowan R, Powers DA. 1992. Ribosomal RNA sequences and the diversity of symbiotic dinoflagellates (zooxanthellae). Proceedings of the Nati onal Academy of Sciences 89 : 3639-3643. Rowan R, Knowlton N, Baker A, Jara J. 1997. Landscape ecology of algal symbionts creates variation in epis odes of coral bleaching. Nature 388: 265-269. Rutzler K. 1990. Association between Caribbean sponge s and photosynthetic organisms. In: New Perspectives in Sponge Biology . (Rutzler K, ed.).Washington, DC: Smithsonian Press. Schneider JA. 1995. Phylogeny of the Cardiidae (Mollu sca Bivalvia): Protocardiinae, Laevicardiinae, Lahilliinae, Tulongocar diinae subfam. n. an d Pleuriocardiinae subfam. n. Zoologica scripta 24(4) : 321-346. Schneider JA. 1998 . Phylogeny of the Cardiidae (Bival via): Phylogenetic relationships and morphological evolution within the subfamilies Clinocardiinae, Lymnocardiinae, and Tridacninae. Malacologia 40(1-2) : 321-373. Schneider JA, Carter JG. 2001. Evolution and phylogene tic significance of Cardioidean shell microstructure (Mollusca: Bivalvia ). Journal of Paleontology 75(3): 607-643. Schneider JA, " Foighil D. 1999. Phylogeny of Giant Clams (Cardiidae: Tridacninae) based on partial mitochondrial 16S rDNA gene sequences. Molecular Phylogenetics and Evolution 13(1): 59-66. Seilacher A. 1990. Aberrations in bivalve evol ution related to photoand chemosymbiosis. Historical Biology 3 : 289-311. Smith SE, Smith FA. 1990. Structure and function of the interfaces in biotrophic symbioses as they related to nutrient transport . New Phytologist 114 : 1-38. Stewart RB. 1930. GabbÂ’s California Cretaceous and Tertiary type lamellibranches. Special Publications of the Academy of Natural Sciences, Philadelphia 3: 1-314, pls 1-17. Suzuki Y, Glazko GV, Nei M. 2002. Overcredibility of mol ecular phylogenies obtained by Bayesian phylogenetics. Proceedings of the National Academy of Sciences 99(25): 16138-16143. Swofford DL. 2002. PAUP* Phylogenetic Analysis Using Parsimony (*and Other Methods). Version 4 . Sunderland, MA: Sinauer Associates. Taylor JD, Glover EA. 1997a. A chemosymbiotic lucinid bivalve (Bivalvia: Lucinoidea) with periostracal pipes: F unctional morphology and description of a new genus and species. In: The marine flora and fauna of the Houtman Abrolhos, Western Australia (Wells FE, ed.). Perth: Western Australian Museum.
115 Taylor JD, Glover EA. 1997b. The lucinid bivalve genus Cardiolucina (Mollusca: Bivalvia, Lucinidae): systematic s, anatomy and relationships. Bulletin of the Natural History Museum, London (Zoology) 63 : 93-122. Taylor JD, Glover EA. 2000. Functional anatomy, chemosymbi osis and evolution of the Lucinidae. In: The Evolutionary Biology of the Bivalvia (Harper EM, Taylor JD, Crame JA eds.). Geological Society of London Special Publications 177: 207225. Taylor JD, Glover EA. 2002. Lamellobranchia : a new genus of lucinid bivalve with four new species form the Indo-West Pacific. Journal of Conchology 37 : 317-336. Telschow A, Hammerstein P, Werren JH. 2005 . The effect of Wolbachia versus genetic incompatibilities on reinforcement and speciation. Evolution 59(8): 16071619. Thacker RW, Starnes S. 2003. Host specificity of the symbiotic cyanobacterium Oscillatoria spongeliae in marine sponges, Dysidea spp. Marine Biology 142 (4): 643-648. Thompson JD, Gibson TJ, Plewniak F, Jeanmougin F, Higgins DG. 1997. The CLUSTALX windows interface: Flexible strategies for multiple sequence alignment aided by quality analysis tools. Nucleic Acids Research 25(24): 48764882. Trench RK. 1979. The cell biology of plant-animal symbiosis. Annual Review of Plant Physiology 30: 485-531. Trench RK, Wethey DS, Porter JW. 1981. Observations on the symbiosis with zooxanthellae among the Tridacnid ae (Mollusca: Bivalvia). Biological Bulletin 161 : 180-198. Umeshita H, Yamasu T. 1985. On the morphology of a speci es of strawberry cockle Fragum sp. The Biological Magazine of Okinawa 23 :50. [In Japanese]. Van Aartsen JJ, Goud J. 2000. European marine Mollusca: Notes on less well-known species. XV. Notes on Lusitanian species of Parvicardium Monterostao, 1884, and Afrocardium richardi (Audouin, 1826)(Bivalvia, Heterodonta, Cardiidae). Basteria 64: 171-186. Veron JE. 2000. Corals of the World . Townsville, AU: Australian Institute of Marine Sciences. Vidal J. 1994. A review of the genus Fulvia Gray 1853 (Mollusca, Bivalvia). Apex 9(4): 93-118.
116 Vidal J 1997a. Large Trachycardiinae from the Indo-West Pacific: The group of Vasticardium orbita (Broderip & Sowerby, 1833) (Mollusca, Cardiidae ). Molluscan Research 18 : 11-32. Vidal J. 1997b. Taxonomic revision of the Indo-Pacific Vasticardium flavum species group. Zoosystema 19(2-3): 233-253. Vidal J. 1999. Taxonomic review of the elongated cockles: Genera Trachycardium , Vasticardium and Acrosterigma (Mollusca, Cardiidae). Zoosystema 21(2): 259335. Vidal J. 2000b. Classification of Cardiidae. Phuket Marine Biological Centre Special Publication 21 (3): 639-644. Vokes HE. 1989. Neogene palaeontology in the northern Dominican Republic, 9. The family Cardiidae (Mollusca, Bivalvia). Bulletins of American Paleontology 97 : 95-161. Voskuil RPA, Onverwagt WPH. 1989. Inventarisation of the Recent European and west African Cardiidae (Mollusca, Bivalvia). Gloria Maris 28 (4&5) : 49-96. Watson ME, Signor PW. 1986. How a clam builds windows: Shell microstructure in Corculum (Bivalvia: Cardiidae). The Veliger 28(4): 348-355. Wheeler WC, Cartwright P, Hayashi CY. 1993. Arthropod phylogeny: a combined approach. Cladistics 9 : 1-39. Williams ST, Taylor JD, Glover EA. 2004. Molecular phylogeny of the Lucinoidea (Bivalvia): Non-monophyly and separate ac quisition of bacterial chemosymbiosis. Journal of Molluscan Studies 70 :187-202. Wilson BR, Stevenson SE. 1977. Cardiidae (Mollusca: Bivalvia) of Western Australia. Western Australian Museum Special Publication 9 : 114 pp. Yamasu T. 1988a. Animals in coral reef 3-Symbiotic relationships in coral reef. In: The coral reefs of Okinawa (Nishimira M ed.). Naha: Okinawaken Kankyokagaku Kensa Centre [In Japanese]. Yamasu T. 1988b. Symbiosis between marine animals and algae. Heredity 42: 12-20 [In Japanese]. Yonge CM. 1936. Mode of life, feeding, digestion a nd symbiosis with zooxanthellae in the Tridacnidae: Great Barrier Reef Expedition, 1928-29, British Museum (Natural History), Scientific Report 1(11) : 283-321. Yonge CM. 1981. Functional morphology and evoluti on in the Tridacn idae (Mollusca: Bivalvia: Cardiaceae). Records of the Australian Museum 33 :735-777.
117 Zann LP. 1980. Living together in the sea . Neptune, NJ: TFH Publications.
118 BIOGRAPHICAL SKETCH Lisa Kirkendale received her B.Sc. from th e University of Victoria, with a focus on marine studies, in 1995. She then went on to obtain a M.Sc. from the University of Guam Marine Laboratory in 1997. She will be conti nuing her research of marine invertebrate molecular phylogenetics at the Un iversity of Sydney, Australia.