GENETIC VARIATION AMONG POPULATIONS AND SPECIES OF GROUPERS AND CORALGROUPERS (OSTEICHTHYES: SERRANIDAE: EPINEPHELINAE) By JOEL L. CARLIN 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 2003
Copyright 2003 by Joel L. Carlin
Dedicated in loving memory of my parents, F. J. Carlin and K. M. Carlin.
ACKNOWLEDGMENTS W. J. Lindberg and B. W. Bowen deserve recognition for mentorship and project development. I thank M. Brenner, C. Cichra, M. Miyamoto, and D. Murie for excellent graduate committee service. J. Acevedo, J. C. Avise, A. L. Bass, L. Bullock, G. Burgess, R. W. Chapman, J. Colborn, D. DeMaria, A. M. Ecklund, A. Francisco, S. Gardieff, T. Gomes, J. Horrocks, S. Johnson, S. A. Karl, T. Maggio, W. O. McMillan, A. Miles, K. Moots, A. Muss, D. R. Robertson, R. H. Robins Jr., C. Rocha, L. Rocha, P. Santos, T. Streelman, D. Weaver and D. Wyanski generously contributed time, expertise, and samples. The Florida Museum of Natural History provided curatorial support; and the Kansas University Natural History Museum loaned tissue samples. Some samples were sequenced by the University of Florida DNA Sequencing Core and/or the Nevada Genomics Center. A portion of this work was funded by a National Science Foundation grant (NSF Award No. 9727048) jointly awarded by the Population Biology and Biological Oceanography Divisions to B. W. Bowen and D. R. Robertson. Field work was also supported in part by the Smithsonian Tropical Research Institute, with logistic support from the US Air Force, the Direcco das Pescas do So Tom, the Administrator of Ascension Island, and the Governor of St. Helena Island. I was funded by the International Womenâ€™s Fishing Association, the Florida Chapter of the American Fisheries Society, the PADI Foundation, PADI Project AWARE, and the UF Department iv
of Fisheries and Aquatic Sciences. All work complied with required animal care and use guidelines. v
TABLE OF CONTENTS page ACKNOWLEDGMENTS.................................................................................................iv LIST OF TABLES...........................................................................................................viii LIST OF FIGURES...........................................................................................................ix ABSTRACT.........................................................................................................................x CHAPTER 1 INTRODUCTION........................................................................................................1 Statement of Purpose...................................................................................................1 Why Study Groupers?..................................................................................................2 How Should Genetic Variation Be Assessed?.............................................................3 Molecular Tools....................................................................................................3 Optimality Criteria................................................................................................4 Species Concepts..................................................................................................6 Gene trees are not species trees......................................................................7 Trees show relationship; not status................................................................7 Operational and theoretical species concepts may not be the same...............7 Summary......................................................................................................................8 2 ANCIENT DIVERGENCES AND RECENT CONNECTIONS IN TWO TROPICAL ATLANTIC REEF FISHES Epinephelus adscensionis AND Rypticus saponaceous (PERCOIDEI: SERRANIDAE)...........................................................10 Introduction................................................................................................................10 Materials and Methods..............................................................................................13 Results........................................................................................................................17 Rock Hind...........................................................................................................17 Greater Soapfish.................................................................................................18 Discussion..................................................................................................................20 Biogeographic Barriers: Caribbean versus Brazil.............................................28 Biogeographic Barriers: West versus East Atlantic..........................................30 vi
3 PHYLOGENETIC RELATIONSHIPS AMONG SOME GROUPER AND CORALGROUPER SPECIES, WITH COMMENTS ON THE TRIBE EPINEPHELINI (PERCOIDEI: SERRANIDAE: EPINEPHELINAE)...................43 Introduction................................................................................................................43 Materials and Methods..............................................................................................47 Results........................................................................................................................52 Cytochrome b.....................................................................................................53 Nuclear Intron.....................................................................................................56 Discussion..................................................................................................................57 Monophyly of the Epinephelini..........................................................................60 Conclusions and Recommendations...................................................................62 4 CONCLUSIONS........................................................................................................76 APPENDIX COMMENTS ON GROUPER SPECIES RELATIONSHIPS....................78 Species Relationships within a Reduced Epinephelini..............................................78 The Hind Clade...................................................................................................78 The True Grouper Clade.....................................................................................79 The Haifa and Leather Bass Clades....................................................................80 The Mycteroperca Clade....................................................................................81 The American Clade...........................................................................................82 The Pacific Grouper and Promicrops Group......................................................83 Final Thoughts...........................................................................................................84 LIST OF REFERENCES...................................................................................................86 BIOGRAPHICAL SKETCH...........................................................................................101 vii
LIST OF TABLES Table page 2-1 Genetic variation in the rock hind Epinephelus adscensionis and greater soapfish Rypticus saponaceous..................................................................33 2-2 Haplotype occurrences and sample sizes in Epinephelus adscensionis.................34 2-3 Population pairwise differentiation observed in Epinephelus adscensionis..........36 2-4 Absolute haplotype frequencies and sample sizes in Rypticus saponaceous........37 2-5 Population pairwise differentiation observed in Rypticus saponaceous................38 3-1 Four alternate classifications of grouper and coralgrouper genera........................66 3-2 Nucleotide composition of DNA sequence alignments from percoid fishes sampled at cytochrome b (cyt b) and LDHA intron 6....................67 3-3 Results of a posteriori topology tests among 50% majority-rule consensus trees generated from epinephelin cytochrome b...................................68 3-4 Results of a posteriori topology tests among 50% majority-rule consensus trees generated from epinephelin LDHA6 introns................................69 3-5 Results of topology tests among unconstrained gene trees and gene trees with a priori topological constraints.....................................................70 viii
LIST OF FIGURES Figure page 2-1 Biogeographic divisions of the tropical Atlantic with sample localities...............39 2-2 Distances among Epinephelus adscensionis haplotypes (Aâ€“HH) and selected congeners, as indicated by a neighbor-joining (NJ) tree...................40 2-3 Distances among Rypticus haplotypes (1) and selected congeners, as indicated by a neighbor-joining (NJ) tree........................................41 2-4 Frequency distributions for pairwise haplotype distances for cytochrome b from Epinephelus adscensionis and Rypticus saponaceous...........42 3-1 Phylogenetic hypotheses of epinephelin evolution................................................71 3-2 The most likely gene tree at cytochrome b as selected by SH tests, showing derived species only...............................................................................................72 3-3 The most likely gene tree at cytochrome b as selected by SH tests, showing ancestral species only.............................................................................................73 3-4 The most likely epinephelin gene tree at LDHA6 as selected by SH tests.............................................................................................................74 3-5 Hypothetical phylogeny of the groupers and their allies.......................................75 ix
Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY GENETIC VARIATION AMONG POPULATIONS AND SPECIES OF GROUPERS AND CORALGROUPERS (OSTEICHTHYES: SERRANIDAE: EPINEPHELINAE) By Joel L. Carlin May 2003 Chair: William J. Lindberg Cochair: Debra J. Murie Major Department: Fisheries and Aquatic Sciences I assessed variability at a portion of the cytochrome b mitochondrial (mtDNA) gene at two taxonomic scales in the groupers and related sea basses (subfamily Epinephelinae, tribe Epinephelini). Intraspecific variation in a grouper (the rock hind Epinephelus adscensionis) and a related outgroup species (the greater soapfish Rypticus saponaceous) clustered into distinct phylogeographic units with boundaries concordant to some biogeographic provinces. The intraspecific study indicated that pelagic larval duration (PLD) and faunal boundaries may only partially determine the limits of gene flow in western Atlantic serranids, and that historical patterns may be important in determining haplotype distribution. To elucidate these historical patterns, a comparison among species of groupers was made at both the cytochrome b gene and a noncoding nuclear intron. Although both mtDNA and nuclear data were concordant in some issues, the LDHA6 intron was not useful in phylogenetic reconstruction of tribal relationships. The x
grouper tribe Epinephelini was not a monophyletic group in the mtDNA gene genealogy, because of the insertion of other families and subfamilies between the coralgroupers (Plectropomus and Variola) and other grouper genera. These data are concordant with other mtDNA loci and with previously described patterns of larval morphology. Taxonomic revisions to the Epinephelini are suggested. xi
CHAPTER 1 INTRODUCTION Statement of Purpose How well does the history of a single gene correlate with described taxonomy, and what does this tell us about the evolutionary relationships of groupers in different parts of the world? Here I describe the genetic variation within and among species of groupers and their allies (fish in the serranid tribe Epinephelini) at a mitochondrial protein-coding gene (cytochrome b) commonly used for elucidating fish population genetics and phylogenetics. I examine this gene at two different scales. First, I describe the distribution and similarity of intraspecific alleles within a single ocean among (presumably) intraspecific demes. Second, I examine similarity and common features of alleles across the entire tropics in order to reconstruct evolutionary relationships among (presumably) noninterbreeding species. Both provide a gene genealogy that may or may not concur with our knowledge of grouper variability as seen in other lines of evidence (e.g., larval dispersal potential, species distribution, larval and adult morphology, other genetic studies). To rigorously and thoroughly compare known taxonomy to the genetic patterns, four supporting questions had to be answered: Is there sufficient genetic variation to construct robust hypotheses of relationship? Are the patterns of genetic variation concordant with previously described morphological, geological, or distributional patterns? Are genetic patterns in groupers concordant with patterns seen in other taxa, and what are the historical and biogeographical implications of these patterns? What are the implications of these data for management and conservation? 1
2 The answers to these questions depend greatly on the two different scales examined (population demographic processes vs. speciation within a family), but both studies involve related study organisms, laboratory techniques, and analytical tools. Why Study Groupers? The tribe Epinephelini (the groupers, coralgroupers, coneys, and lyretails) comprises tropical and subtropical marine fishes that have considerable economic importance and ecological importance as large demersal predators. Several smaller species are popular as aquarium fish. The excellent food quality of the larger groupers supports a variety of fisheries, from cyanide fishing in the Indo-Pacific (IMA Indonesia 2001) to extensive line and trap fisheries in Europe and the Americas (Coleman 2001). The marine commercial catch of groupers for the United States alone has a mean annual worth of nearly $3 million, not including the equal or greater recreational catch (NMFS 2002). Hermaphroditism, group spawning, relatively slow maturation, and low fecundity are common to many grouper species (Domeier and Colin 1997). Therefore, this group is particularly vulnerable to overfishing (Musick et al. 2000, Sadovy 2001). Knowing the genetic variation within populations can lead to inferences about variation in reproductive and recruitment success, some of the most elusive parameters to measure in wild marine populations (Grant and Bowen 1998). One challenge in identifying different genetic phylogroups is the possible ecological differences among alternate forms found in sympatry (Gilles et al. 2000). Data from populations sampled before and after fishing have shown loss of alleles and allozyme heterozygosity in the slow-growing orange roughy (Smith et al. 1991) and successful founder populations of introduced grouper have been documented (Planes and Lecaillon 1998). Estimates of variation at different geographic and taxonomic scales could be applied to models for
3 currently exploited species whose past genetic diversity is not known. Finally, the cryptic variation in some exploited species has hampered fishery data collection in the Indo-West Pacific because of taxonomic confusion (K. Rhodes, U. of Hong Kong, pers. comm.). The management of multispecies reef fish fisheries, the identification of biodiversity â€˜hotspotsâ€™ for conservation, and studies of comparative biology all require a taxonomy that reflects biological history (Collette and Vecchione 1995, Sweijd et al. 2000, Vecchione et al. 2000), but currently no such phylogenetic resolution is available for serranids (Gill and Kemp 2002). The identification of cryptic regional lineages would help quantify the international trade in groupers. For instance, nondestructive surveys of saltwater aquarium species within consumer countries such as the United States could reveal the geographic distribution of unregulated fishing efforts. How Should Genetic Variation Be Assessed? Molecular Tools A variety of tools are able to quantify genetic variation, from measures of protein mobility variation (allozyme electrophoresis) to differential display of gene expression (microarray analysis). Numerous reviews have listed the assumptions, capabilities, and drawbacks of each method (Fral 2000, Haig 1998, Neigel 1997, Sweijd et al. 2000). Hillis et al. (1996) arranged available techniques along spectra of cost and discriminatory power (e.g., for population genetics, higher taxonomy). Discriminatory power hinges on how much variation is present in the examined genetic material or product. Genetic variation is generated by mutation, and retained mutations can be observed among populations and species to generate average mutation rates, which can vary widely among types of organism and types of marker (Hewitt 2001).
4 Thus experimental design in phylogeography and molecular systematics makes assumptions about available variance in the taxa at a certain genetic locus and seeks to balance these two factors: The best laboratory method capable of measuring the variation. The cost of analyzing the sample size necessary to address the question of interest. Most of the molecular data generated for the study of grouper populations and species were from DNA sequences of the mitochondrial cytochrome b gene. The relationship between mutation rate and statistical power means that questions addressed by this technique at this locus may not be answerable by other loci and techniques; likewise other techniques can expand or refine aspects of this study. Optimality Criteria Discrimination and classification of compared units involves selecting philosophical approaches to guide the data analysis. Many approaches operate under an assumption (implicitly or explicitly) that mutation is rare; and therefore that complex mutational patterns are less likely than simpler ones (Occamâ€™s razor). Each philosophical approach has algorithms that generate numerous possible phylogenetic trees and then attempt to identify the best tree. The analytical approaches to phylogenetics differ in how they measure a treeâ€™s optimality. Molecular phylogenies typically are constructed using one of four optimality criteria (all of which were used to some extent here): distance, parsimony, likelihood, and posterior probability methods. Distance methods optimize the shortest tree as measured directly by some metric, most simply by fit of taxa to overall similarity in the number of nucleotide substitutions. Distance measures therefore assume that similarity among individuals is not due to chance, and that similar units are historically related and
5 not the result of evolutionary convergence. Parsimony methods minimize this assumption by not considering all data. Instead, parsimony optimizes the shortest number of changes, but only considers those features shared by two or more taxa. Likelihood methods use the frequency of nucleotide differences (and lack of differences) to calculate the probability that the observed data would generate a given phylogenetic pattern. Likelihood methods can emphasize both differentiation and â€˜sameness,â€™ and produce results easily evaluated by statistical testing. Posterior probability (Bayesian) methods also use likelihood calculations, but instead of evaluating a modelâ€™s ability to fit observed data, Bayesian methods incorporate prior knowledge of mutation rate and evolutionary history into the likelihood calculations, resulting in a tree with a probability that a relationship could be the result of a model given any data, observed or not. Optimality criteria also differ in sensitivity to an uneven distribution of variance. The lack of close relatives to very divergent individuals can result in highly differentiated taxa being lumped together in a false sister relationship. This long branch attraction results when an algorithm, unable to cluster two extremely different taxa to any other lineages, clusters the divergent pair together whether or not they truly share a common history. The best way to overcome long branch attraction is to include the true relatives (both extant and extinct) of the problem taxa, requiring foreknowledge of the true phylogeny. Unable to sample all described species of groupers worldwide, I searched for criteria that would be least affected by the long branch attraction problem. Studies of this problem by Takahashi and Nei (2000) and my own preliminary analyses (Carlin 2002) were inconclusive as to which criterion and algorithm performed best for the amount of data available.
6 Chapter 2 addresses the null hypothesis that two serranid species each comprise a single freely interbreeding population. If these species are regulated as a cohesive unit, then it is doubtful that similarity between two individuals is due to chance convergence between genetically independent lines. Therefore, I make the a priori choice of distance methodology as the primary optimality criterion. In case the distance assumption of minimal convergence is incorrect, I also conduct a parsimony analysis. Chapter 3 addresses the genetic history of species and genera of groupers and their relatives. I have the a priori expectation that these genetic patterns might be the result of millions, if not tens of millions, of years of divergent evolution. Therefore, I consider distance methods inappropriate for these data. My anticipation of a problem with long branch attraction meant that I considered multiple optimality criteria. I compare results from various parsimony methods to posterior probability analyses, under the assumption that real relationships are more likely to be identified under all criteria. Thus the phylogenetic trees presented are the simplest patterns as defined by various criteria. I realize, however, that the simplest explanation is not necessarily the most correct. Therefore I examined other lines of evidence, such as biogeographic provinces (Chapter 2); or phylogenetic trees from other genetic loci (Chapter 3). In both studies I endeavor to give alternate interpretations of the data, including the placement of species boundaries. Species Concepts The optimality criteria and their associated algorithms are used to identify, but not describe, historical relationships. I propose that delimiting evolutionary relationships is the purview of phylogenetics, while the description of organismal sets is a question of taxonomy. I have used three guidelines to discuss the results.
7 Gene trees are not species trees A cytochrome b genealogy shows the apparent relationship among lineages sampled from single individuals. The genealogy is therefore a reconstruction of observed results, interpreting this strictly as a historical record for all individuals throughout some period of geologic time changes the reconstruction into a phylogenetic hypothesis. The investigator must optimize the sampling design to make this assumption believable. Trees show relationship; not status Single gene genealogies can provide important information even if the status of the member species is in question. If two lineages do not seem to exchange alleles across a proposed barrier, then there is support for the barrierâ€™s existence. If several lineages seem to have recently evolved in the Caribbean, then there is support for the Caribbean as a promoter of genetic diversification. Operational and theoretical species concepts may not be the same Some species concepts describe the act of speciation, and some describe the criterion for identifying species in the field or laboratory. Most attempt to do both, but the criteria for identifying a species as distinct (an operational species concept) is based upon a model of evolution (the theoretical species concept). Most morphological and genetic surveys of tropical reef fishes (of which the groupers are a part) and large reef invertebrates have revealed that the most diverse free-swimming families have all experienced an explosion of speciation in the past few million years. The resulting genetic variation within and among marine fauna tends to be a collection of very ancient and very recent lineages in a patchy geographic mosaic. Norris (2000) showed several mechanisms of speciation in the pelagic environment that were invoked to explain the patchy mosaic of diversity in the sea (e.g., Lessios et al.
8 2001, Palumbi 1996) not explained by simple tectonic boundaries. Such a complex pattern is a reasonable expectation in the groupers. Some groupers hybridize among species (Glamuzina et al. 2001, van Herwerden et al. 2002) and genera (Bostrom et al. 2000), while there is an apparent lack of interbreeding between cryptic units within other grouper species (Gilles et al. 2000). If grouper diversification is very recent, then the ancestral units of some species may possibly be the same modern species that can be seen today, meaning that the ancestral species should not be regarded as a separate evolutionary unit. In contrast, ancient fish lineages are often found in remote areas or with highly specialized life histories; and these may represent apparent anagenetic change (i.e., evolutionary change without lineage splitting). I view speciation in dispersive marine fauna (i.e., not gobies or blennies) to be best described by an evolutionary species concept, in which cohesive lineages through time may split off or coalesce. While the evolutionary species concept shows a view of evolution that may (or may not) be accurate for the groupers, the operationality of the evolutionary species concept is made difficult by its predictive nature and the lack of a fossil record for groupers. An operational concept that accepts limited hybridization among lineages, recognizes persistent ancestors, and has the ability to discriminate recently evolved units is the PSC-1 variant (Brooks and McLennan 2002) of the phylogenetic species concept (Cracraft 1983, Cracraft 1987). However, because gene trees represent limited sampling of evolutionary history, I revised taxonomy only when supported by other lines of evidence. Summary Within this dissertation, I elucidate evolutionary relationships in a large set of widely distributed and highly variable warmwater marine fishes by comparing observed
9 cytochrome b gene genealogies to information from morphology, distribution, and other loci. In Chapter 2, my dissertation examines variation of cytochrome b to describe gene flow limits in the recent ecological past among populations of the rock hind (the grouper Epinephelus adscensionis) and the greater soapfish (Rypticus saponaceous, in a sister subfamily to groupers). The variation (or lack thereof) in cytochrome b at this scale is compared to nongenetic evidence from the groupers (their pelagic larval dispersal ability) and from many fish families (described zoogeographic boundaries). My sampling design is optimized to address whether or not each of these species are monophyletic units (i.e., a single cohesive lineage distinct from other taxa) with free gene flow (i.e., panmixia) across their geographic distributions. In addition to reconstructing relationships among individuals and localities, I use an analysis of molecular variance test to determine if different alleles are evenly distributed, as predicted by a panmictic model. Secondly, I use the same locus to examine variation across multiple grouper species to build a phylogenetic hypothesis (evolutionary tree) that describes evolution in the groupers. My sampling is optimized to address whether the grouper tribe (Epinephelini) is a monophyletic unit. As in the Chapter 2 phylogeographic study, I compare the cytochrome b genealogy to other lines of evidence, including a nuclear noncoding region and previous hypotheses derived from morphology and a different mitochondrial gene. Finally, the data from both studies should expand our knowledge of gene flow in groupers, contributing to identifying management units and delimiting areas of biodiversity concentration.
CHAPTER 2 ANCIENT DIVERGENCES AND RECENT CONNECTIONS IN TWO TROPICAL ATLANTIC REEF FISHES EPINEPHELUS ADSCENSIONIS AND RYPTICUS SAPONACEOUS (PERCOIDEI: SERRANIDAE). Introduction Studies of population genetic structure in marine organisms often reveal low levels of differentiation (Bucklin and Wiebe 1998, Graves 1998, Ward et al. 1994). In reef fishes, which are typically nonmigratory as juveniles and adults (Leis 1991), little or no genetic population structure has been observed among widely separated localities in the Caribbean Sea. Shulman and Bermingham (1995) examined mitochondrial DNA (mtDNA) from eight Caribbean species. Lacson (1992) surveyed 33 allozyme loci from five Caribbean species. Collectively these studies included six taxonomic families exhibiting a wide range of pelagic larval durations (from approximately 13 to 122 days). Estimates of divergence among localities within the Caribbean Sea and Gulf of Mexico were low (ST = 0.000 to 0.172, Shulman and Bermingham 1995; ST = 0.004 to 0.035, Lacson 1992). On the same geographic scale, an allozyme survey of the damselfish Stegastes partitus showed no significant population structure (Lacson 1992, Lacson et al. 1989). The lack of divergence between distant localities is usually attributed to egg and larval dispersal over hundreds or several thousand kilometers (see Brothers and Thresher 1985, Johannes 1978), a mechanism invoked to explain the presence of reef fish species on isolated islands (e.g. Allen and Robertson 1997, Robertson 2001). Shulman and Bermingham (1995: 908) concluded that â€œunderstanding dispersal and consequent 10
11 genetic differentiation among populations will require that we direct our attentions to the capabilities and behaviors of larvae that result in limitation or augmentation of dispersal distances.â€ Briggs (1995) described two distinct biogeographic provinces within the tropical northwest Atlantic: a West Indian Province containing the islands from Cuba and the Bahamas to Grenada and the isolated northerly island of Bermuda (Figure 2-1). The West Indies fauna are distinct from those of the Caribbean biogeographic province, which includes southern Florida, the southern Gulf of Mexico, and the coasts of Central America and northern South America. Because these provinces are separated by as little as 60 km, Briggs (1995) proposed that strong, fast currents could act as vicariant barriers at the Florida Straits in the north, the Yucatn Channel in the west, and between Trinidad and Grenada in the south. However, to date, genetic studies have not indicated a restriction of gene flow between the Antilles and continental locations (Rocha et al. 2002, Shulman and Bermingham 1995). If a reef area as large as the Caribbean Sea and Gulf of Mexico can be connected by contemporary gene flow, then at what scale can one expect to find genetic isolation of reef populations? Many of the reef species surveyed in the Caribbean also occur in other biogeographic provinces of the Atlantic. If multiple biogeographic provinces are considered, evolutionary genetic partitions could be detectable despite a relatively lengthy pelagic larval duration. Significant divergences have been observed between biogeographic provinces in antitropical species (e.g., Ball et al. 2000, Bowen and Grant 1997); and tropical species may also show genetic structure between discontinuous
12 biogeographic zones (Banford et al. 1999, Colborn et al. 2001, Muss et al. 2001, Planes et al. 1993, Rocha et al. 2002). For Atlantic reef species, all of the major biogeographic provinces are discontinuous (Briggs 1995). Coral reefs in the Gulf of Guinea (West Africa) are separated from Brazilian counterparts by an oceanic gap of 4,000 km, a distance postulated to be an effective barrier to dispersal (Rosen 1976). Caribbean reefs are separated from Brazilian reefs by 2,300 km of coastline dominated by soft bottoms and riverine outflows from the Amazon and Orinoco rivers. These outflows form coastally trapped discharge zones of hyposaline water (Bonilla et al. 1993, Longhurst 1998), potentially decreasing effective latitudinal dispersal in the tropical western Atlantic. These barriers are invoked to explain the high endemism of the Brazilian and Caribbean provinces, and this level of isolation may be higher than is currently recognized. Recent research indicates a potentially large number of undescribed endemic reef fish species in the Brazilian province (Bernardi et al. 2000, Floeter and Gasparini 2000, Floeter et al. 2001, Rocha and Rosa 1999), and this finding also may be applicable to the poorly studied reefs in West Africa. Vicariant barriers between Atlantic reef provinces may be sufficient to allow allopatric speciation in some cases (Bernardi et al. 2000, Muss et al. 2001). In other cases, there has clearly been some dispersal, gene flow, or colonization in recent evolutionary time, as evidenced by tropical Atlantic reef fish species that occur on both sides of the Atlantic. One possible mechanism for mediating gene flow between the Brazilian and East Atlantic biogeographic provinces in particular is the presence of oceanic islands (e.g., Ascension, Saint Helena, and Trindade Island) (Figure 2-1). While
13 small in size, the islands that interrupt the deep mid-Atlantic expanse may serve as stepping stones for transoceanic colonization (Briggs 1995). With these issues in mind, the genetic variation in two amphi-Atlantic reef fishes was surveyed. The rock hind, Epinephelus adscensionis Osbeck (1971), is a serranid fish that occurs in shallow water from Bermuda to Brazil, on the mid-Atlantic ridge islands, and in the eastern Atlantic from the Gulf of Guinea to Morocco (Bhlke and Chaplin 1993). The greater soapfish, Rypticus saponaceous Bloch and Schneider (1801), is another amphi-Atlantic serranid species (as defined by Guimares 1999) that has a distribution similar to the rock hind. Both species are demersal reef fishes (Thresher 1980), such that long-range dispersal is limited to the pelagic larval stage (Leis 1991). The pelagic larval duration of rock hind and greater soapfish have not been reported, but a 40-day average is proposed for serranids (Lindeman et al. 2000). If barriers to dispersal by pelagic larvae exist in these amphi-Atlantic reef fishes, then sustained interruptions in gene flow may be evident on a broad geographic scale. Molecular genetic analysis of rock hind and greater soapfish from all major biogeographic zones within their ranges should therefore elucidate the population structure in these reef fishes; and may illuminate general biogeographic patterns for Atlantic reef species. Materials and Methods Samples of E. adscensionis, R. saponaceous, and selected congenerics were collected from multiple locations within the Caribbean, Brazilian, and mid-Atlantic biogeographic provinces, and eastern Atlantic region (Figure 2-1). Tropical northwest Atlantic specimens of E. adscensionis were collected from South Carolina, the Florida Keys, the Bahamas and Barbados. Brazilian rock hind specimens were collected from Trindade Island and the mainland states of Vitoria and Bahia. Mid-Atlantic ridge
14 collections were made at Ascension Island and Saint Helena. Eastern Atlantic E. adscensionis were obtained from So Tom (Gulf of Guinea). Specimens of R. saponaceous were obtained from Grenada and Barbados (tropical NW Atlantic); Vitoria and Bahia (Brazilian province); Ascension and Saint Helena (mid-Atlantic ridge province); and the eastern Atlantic localities of So Tom and Cape Verde. Fish were collected during snorkel and SCUBA diving (using polespears and microspears) primarily by D. R. Robertson and B. W. Bowen. In most locations, gill and/or muscle samples were taken and stored in a saturated-salt DMSO buffer (Amos and Hoelzel 1991). Total genomic DNA was isolated with organic extraction (using phenol, chloroform, and isoamyl alcohol) and precipitated with sodium acetate in a 95% ethanol solution. Isolated DNA was resuspended in 100 L TE (10 mM Tris and 1 mM EDTA, pH 8.0). An approximately 750 base pair (bp) fragment of the mitochondrial cytochrome b gene was amplified via the polymerase chain reaction (PCR; Saiki et al. 1985) with primers Cyb-09H (5'-GTGACTTGAAAAACCACCGTTG-3', Song et al. 1994) and Cyb-07L (5'-AATAGGAAGTATCATTCGGGTTTGATG-3', Taberlet et al. 1992). The amplification reaction mix included 3 mM MgCl, 40 nM of each primer, 17.5 nM of each dNTP, 0.80 L of Taq DNA polymerase (Promega, Inc., Madison WI) and 3.5 L of 10x PCR buffer (Promega, Inc.) in 35 L total volume. Amplification temperature regime included an initial 3-minute denaturation at 94C and final extension of 72C for three min, with 25 cycles of 30 sec at 94C, 30 sec at 53C, and 40 sec at 72C for E. adscensionis. Amplifications of R. saponaceous DNA used an identical protocol, except for an annealing temperature of 54C and a reaction buffer with 1 mM MgCl2.
15 Oligonucleotide primers were removed by Ultrafree-MC centrifugal filter units (Millipore Corp., Billerica MA) and/or simultaneous digestion with exonuclease I and shrimp alkaline phosphatase (USB Corp., Cleveland OH). Amplification products were denatured and DNA sequencing reactions were conducted with fluorescently labeled dideoxy terminators according to manufacturer's recommendations (ABI Model 800, Applied Biosystems, Inc., Foster City, CA). Labeled extension products were gel separated and analyzed with an automated sequencer (ABI Models 373A and 377) in the University of Florida DNA Sequencing Core. Resulting fragments were aligned and edited using Sequencher ver. 3.0 (Gene Codes Corp., Ann Arbor, MI). All samples were sequenced in forward direction using Cyb-09H; or in the reverse direction using Cyb-19H (5'-CTCACTGCTCGGACTCTG-3') for some samples of E. adscensionis or using Cyb-22H (5'-GTGAARTTGTCTGGGTCTCCT-3') for some samples of R. saponaceous. Both forward and reverse sequences were obtained for at least one representative sample for each haplotype, to assure accuracy of haplotype assignments. Sequences were aligned with representative species of Epinephelus (20 spp.) and Rypticus (2 spp.) to confirm identification and provide outgroup comparisons. Haplotypes were also compared with the complete cytochrome b sequences from Homo sapiens (GenBank Accession V00662, Anderson et al. 1981) and Epinephelus sp. A (GenBank Accession AF143193, Allegrucci et al. 1999) as additional quality controls. For clarity of discussion, rock hind haplotypes are designated by letter abbreviations, while numerals are used for greater soapfish haplotypes. Population genetic parameters were calculated with Arlequin ver. 2.0 (Schneider et al. 2000). Nucleotide diversity (; Equation 10.5 in Nei 1987) and haplotype diversity
16 (h; Equation 8.5 in Nei 1987) were calculated for each biogeographic province. Genetic distances (d) were calculated using the substitution model of Tamura and Nei (1993) corrected for unequal mutation rate ( = 0.50; Yang 1996) with a 3:1 transition to transversion (ti:tv) ratio. The distribution of pairwise differences among haplotypes generated a mismatch distribution (Rogers 1995, Rogers and Harpending 1992), with the sum of squared differences in frequency at each mismatch class, or raggedness index (r; Harpending 1994). A resampling simulation (100 bootstraps) generated an expected distribution of pairwise differences using population-expansion parameters estimated by a least squares nonlinear regression (Schneider and Excoffier 1999). The proportion of genetic diversity within and among localities was estimated with an analysis of molecular variance (AMOVA) in the Arlequin ver. 2.000 software package (Excoffier et al. 1992). The total variance explained by population-level differences (ST) was estimated to infer degree of population subdivision. A population pairwise differentiation test was performed using a 10,000-step Markov chain simulation (Schneider et al. 2000). Significance (overall = 0.05) was determined after full Bonferroni correction for multiple comparisons (individual P = 0.002 for rock hind and P = 0.0034 for greater soapfish). Both majority-rule Tamura-Nei neighbor-joining (NJ) and maximum parsimony (MP) trees were constructed and evaluated using 500 bootstrap replicates. The NJ tree includes two congeners: Epinephelus analogus (n = 3) and E. labriformis (Figure 2-2). Rather than reporting only the most parsimonious phylogram(s), a consensus tree considering slightly less parsimonious solutions was constructed. This was reported only if there were few alternate solutions (i.e., < 25 trees < 5 steps longer than the most
17 parsimonious cladogram). All phylograms were constructed using Paup* ver. 4.0b (Swofford 2002). Results Rock Hind A total of 109 E. adscensionis were surveyed, yielding 593 bp of resolved sequence per individual. No extra termination codons were observed in comparison with the Allegrucci et al. (1999) serranid sequence. Sequence comparisons revealed a transition-rich data set (ti:tv ratio = 23:5) that included a total of 34 haplotypes (d = 0.002 to 0.044; Tables 2-1, 2-2). However, only seven haplotypes were observed in more than one specimen. These sequences revealed low nucleotide diversity ( = 0.008) and moderate haplotype diversity (h = 0.586), a common pattern in marine fishes (Grant and Bowen 1998). The AMOVA indicated very strong population structure for E. adscensionis (ST = 0.867, P < 0.001). Differentiation tests indicated significant population structure, with pairwise STâ€™s > 0.9 between Florida and all other localities except for comparisons with South Carolina and Trindade Island, for which low sample sizes (n = 3 and 2, respectively) precluded testing (Table 2-3). The population structure is characterized by deep divergence between the southeastern United States haplotypes and those of all other locations and the abundance of haplotype A (64.2% of all rock hind samples) outside Florida/South Carolina. The mean sequence divergence between Florida/South Carolina haplotypes and all other haplotypes was d = 0.040 (range: 0.034 to 0.044). Removing the highly divergent Florida/South Carolina samples from AMOVA decreased diversity measures of the remaining samples (h = 0.481, = 0.001, ST = 0.056, P < 0.001). Significant differentiation was also found in comparing the Bahamas sample with the
18 Ascension and St. Helena samples (Table 2-3), possibly reflecting the predominance of haplotype A in the mid-Atlantic localities, in contrast to its absence in available Bahamas specimens (n = 5) (Table 2-2). The MP tree had the same structure as the NJ tree: two primary branches in E. adscensionis with bootstrap values 0.90. All equally parsimonious trees had unresolved polytomies among those haplotypes found together in only one population (and usually differing by only one mutation). The close relationship among haplotypes within each branch of the rock hind phylogeny also created a bimodal mismatch distribution (r = 0.070, q0 = 1.169, q1 = 1.169). When a distribution was estimated after removal of the Florida/South Carolina samples, no significant difference was found between the observed haplotype distribution and an equilibrium distribution simulated under a rapid population expansion model ( = 1.126, P = 0.550). Aside from haplotype A, only haplotype V (one specimen each from Barbados and St. Helena, 6,745 km apart), was found in multiple biogeographic provinces. The Caribbean province, however, is split by the presence of 21 to 29 nucleotide differences between haplotypes D to I from the Florida Keys and haplotypes A and J to V in the Bahamas and Barbados. Greater Soapfish A total of 86 greater soapfish specimens were analyzed, including 27 specimens from the East Atlantic Region (Cape Verde and So Tom) that diverged notably from samples in the remainder of the range (d = 0.044, range: 0.037 to 0.052). These fish matched conventional descriptions of R. saponaceous (Fischer et al. 1981) and were not consistent with the only other congeneric in the East Atlantic (Guimares 1999). Analyses of 682 bp of cytochrome b yielded a conventional pattern of mutations
19 (ti:tv = 27:6) and a total of 31 R. saponaceous haplotypes (Tables 2-1, 2-4) with somewhat higher genetic diversity ( = 0.019, h = 0.929) than the rock hind ( = 0.008, h = 0.586). As in rock hind, the dataset was influenced strongly by a bifurcation between samples in one portion of the range (eastern Atlantic for this species) and a branch containing all other collection sites (western and central Atlantic). The divergence between the eastern Atlantic haplotypes and all others yielded strong population structure (ST = 0.876, P < 0.001). Large pairwise STâ€™s were calculated in comparisons between Cape Verde or So Tom with all other localities (ST = 0.906 to 0.963, Table 2-5); all but one of these comparisons were significant differentiation tests after Bonferroni correction ( = 0.003). Reanalysis of the R. saponaceous haplotypes, after excluding the divergent East Atlantic samples, reveals significant overall population structure (ST = 0.372, P < 0.001). Nineteen specimens (22% of the total) possessed a haplotype unique within this study; haplotypes 1 and 5 are the only ones present in more than one biogeographic province. Within the West Atlantic samples, the low degree of haplotype sharing resulted in significantly different pairwise differentiation in comparisons between Brazil versus Ascension, Brazil versus St. Helena and Grenada versus St. Helena (ST = 0.327 to 0.695, P < 0.001; Table 2-5). To examine the influence of small sample size in the Caribbean localities, an additional analysis was conducted considering the Barbados and Grenada as a single population (h = 0.917; = 0.003; overall ST = 0.886, P < 0.001; West Atlantic overall ST = 0.354, P < 0.001). In this case only the comparison between the combined Caribbean sample and So Tom changed in significance (ST = 0.927, P = 0.002). The distribution of observed differences between all pairs of haplotypes (r =
20 0.194, q0 = 0.000, q1 = 2.646) was significantly different from the equilibrium model of rapid population expansion ( = 4.430, P 0.49). Analysis of three Rypticus species available in this study indicates that the two lineages of putative R. saponaceous described here may not be monophyletic (67% bootstrap support), and indeed form an unresolved polytomy with five specimens (three haplotypes) of Rypticus bicolor, an eastern Pacific species, and R. subbifrenatus, another amphi-Atlantic species. Mean interspecific (R. saponaceous vs. R. bicolor) pairwise sequence divergence was d = 0.047 (range: 0.043 to 0.055). The only other soapfish reported from the eastern Atlantic, R. subbifrenatus, was collected at the same locality in So Tom as the putative R. saponaceous specimens. The mean genetic distance between R. saponaceous and R. subbifrenatus was d = 0.115 (range: 0.112 to 0.124), greater than that found between the East Atlantic and other samples of R. saponaceous (d = 0.044). Discussion Several patterns concerning the phylogeographic histories of these two species can be recovered from these data. First, both rock hind and greater soapfish intraspecific phylogenies are characterized by a single major bifurcation, with one branch restricted to a single region while the other is widespread through multiple biogeographic provinces. Haplotypes within the divergent branches exhibit a shallow history and rapid coalescence to a common ancestor. For the rock hind, a single haplotype was shared over most of the speciesâ€™ distribution, and the large number of â€œtwigsâ€ is consistent with a population bottleneck and subsequent expansion from two ancestral (and divergent) lineages (Avise 2000). In contrast, greater soapfish exhibited less haplotype sharing among
21 biogeographic provinces, and a null model of rapid population expansion was not consistent with the structure of the western and central Atlantic lineages. A reconstruction of the biogeographic and evolutionary history in greater soapfish and rock hind is attempted by first considering the prominent, relatively ancient, separations between phylogroups. The greatest interlocality divergences found in rock hind and greater soapfish (d = 4.4% and 5.2%, respectively) are relatively deep for intraspecific comparisons of marine fishes (see reviews by Avise 2000, Grant and Bowen 1998). However, descriptions of relatively ancient cytochrome b divergences in marine fishes are accumulating. Bernardi et al. (2000) observed a divergence of d = 2.3% in the parrotfish Sparisoma rubripinne between two eastern Atlantic (So Tom) and two Caribbean specimens. Divergences of d = 0.052 to 0.127 were observed between five phylogroups in biogeographic provinces for the amphi-Atlantic reef blenny Ophioblennius atlanticus (Muss et al. 2001). The deep splits observed in rock hind, greater soapfish, parrotfish, and blenny are of an equivalent magnitude to the divergences observed in cytochrome b between Newfoundland and Norwegian samples of the capelin Mallotus vilosus (d = 5.7%; Birt et al. 1995). These values approach divergences between sister species of other marine fishes (Johns and Avise 1998), but are less than that described between cryptic sympatric species in another grouper, Epinephelus marginatus (d = 0.158; Gilles et al. 2000). The deep divergences observed in E. adscensionis and R. saponaceous also invoke the possibility of cryptic species. This may be particularly likely for the greater soapfish, based on a Pliocene divergence between eastern and western Atlantic specimens. The molecular evolutionary rate of cytochrome b has been estimated to range from 1 to 3
22 %/MY (Canatore et al. 1994, Fleischer et al. 1998, Irwin et al. 1991). Although rates may vary among taxa (Mindell and Thacker 1996), a molecular clock rate of 1 to 2 %/MY is applied. This would place the earliest divergence date between Atlantic R. saponaceous and the East Pacific R. bicolor to 3.5 MYA (range = 2.1 to 5.5 MYA), a timeframe concordant with the last contact between East Pacific and West Atlantic fishes (about 3.1 to 3.5 MY ago; Coates and Obando 1996). This benchmark is used to estimate other divergence dates in the evolutionary history of soapfishes, and provisionally apply this rate to rock hind, with the recognition that divergence dates for these species are not calibrated and must be interpreted with caution. Based on the molecular clock rate of 1 to 2%/MY, the deepest mtDNA lineages in E. adscensionis coalesce at approximately 2.7 MY, and the deepest lineages in R. saponaceous coalesce at about 3.3 MY. Evidence from oxygen isotope ratios in fossil foraminifera indicates the shoaling and separation of waters about the Isthmus of Panama from 5 to 8 MYA, even before the complete emergence of present-day Panama (Collins et al. 1996, Keigwin 1982). The resulting hydrostatic imbalance caused the major current systems of the Atlantic to increase in velocity, then slowed to present levels by ~ 2.5 MY BP to produce the general circulation patterns that persist today (Haug and Tiedemann 1998, Maier-Reimer et al. 1990). Thus, it is possible that divergence occurred among Atlantic Rypticus populations following the closure of the central American isthmus (and subsequent responses in Atlantic circulation) due to decrease in dispersal opportunities across the tropical Atlantic. An obvious alternative is that the specimens collected in Cape Verde and So Tom were misidentified members of another previously described Rypticus species. This
23 possibility is unlikely for three reasons: first, the collection protocol was designed to distinguish R. saponaceous from similar congeners. Specimens of R. saponaceous in the West Atlantic have been collected syntopically with R. bistrispinus (Guimares 1999); these species were differentiated according to coloration following the criteria of Smith (1997). East Atlantic collections included specimens of the only other soapfish described from this region (R. subbifrenatus; Courtenay 1970). Second, the phylogenetic analyses include R. subbifrenatus collected at So Tom and show that it is highly divergent from both eastern and western Atlantic R. saponaceous as well as the East Pacific R. bicolor. Third, Cape Verde and So Tom specimens had no gross departures from prior descriptions of R. saponaceous, the species with the greatest meristic variability in the genus (Courtenay 1967, Guimares 1999). The molecular evidence, in combination with the field observation and geographic considerations, indicate the strong possibility of two species within the currently recognized R. saponaceous. Regardless of taxonomic considerations, the cytochrome b gene tree of greater soapfishes indicates that eastern and western Atlantic lineages initially diverged at about the time of the closure of the Isthmus of Panama. Subsequently, or at nearly the same time (Figure 2-4), the ancestor of R. bicolor (in the east Pacific) diverged from R. saponaceous (in the West Atlantic). This finding invokes the East Atlanticâ€”East Pacific generalized track proposed by Rosen (1976) to explain the distribution of shorefishes. Under Rosenâ€™s hypothesis, the greatest vicariant separation in this region should be between East and West Atlantic, based on the opening of the Atlantic some 80 MY ago. Subsequently, a more recent vicariant separation occurred between the West Atlantic and East Pacific, based on the closure of a central American seaway. Some aspects of fish
24 phylogeography can support this theory (see Banford et al. 1999), and the bicolor/saponaceous split does seem to fit the closure of the West Atlantic/East Pacific connection. However, the East-West split in Atlantic greater soapfishes estimated at 3.3 MY is much too shallow to be based on sea floor spreading. For this aspect of greater soapfish phylogeography, dispersal in an accelerated circulation system (see above) is a more likely explanation. A similar divergence was observed among rock hind, with one mtDNA lineage restricted to the northwest Caribbean (Florida Keys) and the southeast United States. The other lineages were observed throughout the tropical Atlantic, and most notably at two Caribbean collecting sites (Barbados and the Bahamas). How can one explain this curious finding of a major genetic separation of lineages separated by <100km of open water? As above, the possibility of misidentification must be considered. This is especially pertinent for rock hind, because there are several similar congeners in the Caribbean. However, misidentification is unlikely for two reasons: First, the samples were collected by personnel specifically trained to distinguish rock hind from other epinephelines. Second, the Florida/South Carolina specimens were collected by several expeditions and by different trained persons and it is therefore unlikely that all parties collected the wrong species at 100% frequency. What biogeographic process could explain the divergent mtDNA lineages within the tropical West Atlantic rock hind? No obvious differences in size, coloration, or external morphology among rock hind specimens were observed, so it is possible that what is currently called one species consists of separate species: an insular species found in the eastern part of the greater Caribbean and a continental species with a currently
25 unknown range. As in soapfish, the lineages appear to be allopatric, although Florida and South Carolina sample sizes are small. Fast-flowing currents in the Florida Straits have been proposed as vicariant barriers between the West Indian biogeographic province (including the Bahamas) and a Caribbean province containing the Florida Keys (Figure 2-1; Briggs 1995). However, this strong divergence among rock hind is exceptional, because all the prior surveys of Caribbean Sea reef fishes revealed low or no population structure across such distances. In particular, Shulman and Bermingham (1995) showed that, for each of eight reef fish species (not including rock hind or soapfish), populations in Barbados and Florida shared haplotypes both with each other and with other continental and island locations in the Caribbean. Another alternative is a long-term divergence between the divergent Caribbean lineages and those in the Brazilian province (as observed in other reef species; Bernardi et al. 2000, Muss et al. 2001), followed by recent dispersal into the West Indies from Brazil. Grenada, as one of the southernmost outposts of reef habitat in the Caribbean, is suitably located to encounter larvae drifting north on the Brazilian Current. Based on species distributions, geography, and ocean circulation patterns, Grenada is one of the gateways for occasional dispersal between Brazil and the Caribbean (Rocha et al. 2002). The genetic data support this hypothesis: while not found in abundance (n = 7), greater soapfish haplotype 1 was found in specimens from Grenada, Barbados, and Brazil; and rock hind haplotype A was observed in both Brazil and Barbados. Genetic data from large population samples of both species at many sites scattered throughout the tropical northwest Atlantic will be needed to (i) define the distributions of the two main lineages of rock hind and show how they relate to insular/continental biogeographic subdivisions
26 of that region, (ii) show whether those rock hind lineages are allopatric, and (iii) show how the phylogeography of the greater soapfish within the Caribbean relates to that of the rock hind. As the latter species were sampled only in the southeastern Caribbean this study cannot say whether there are lineage subdivisions in that species within the tropical NW Atlantic. Bowen et al. (2001) demonstrated that trumpetfish phylogroups, separated by geography for 3 to 4 MY, hybridized in renewed contact in the South Atlantic. As noted by Knowlton (2000), reproductive isolation in sympatry is an acid test for speciation, and rock hind may be in the early stages of this test. The reproductive compatibility of the two main lineages of both rock hind and greater soapfish may determine whether rock hind have speciated in isolation, or whether this is a single species characterized by deep divergence and subsequent admixture (see Veron 1995). Against the backdrop of strong divergence, both species exhibit many slight differentiations (often by a single transition) within the major lineages. The distribution of these lineages among western and central Atlantic greater soapfish yields significant population genetic structure, but this is not the case in rock hind. The relative abundance of a widespread haplotype among biogeographic provinces and regions (as seen in rock hind haplotype A), accompanied by many closely related haplotypes, yielded a low level of population structure in rock hind samples from the Bahamas to So Tom. The lack of heterogeneity in rock hind may be due to extensive gene flow between biogeographic provinces and regions. Alternately, the shallow gene genealogy with widespread haplotype sharing could arise from selection or demographic processes, working either singly or in combination (Grant and Bowen 1998). Such demographic processes include
27 the effects of temporal variation in recruitment success and large fluctuations in population size. A shallow but diverse gene genealogy can be generated by extremes in recruitment success (Hedgecock 1994), where all recruits to a year-class were generated by a few successful mating events. While the population may have a large number of spawners, in sweepstakes reproduction successful cohorts of recruits are drawn from only the few propagules that meet a specific and stringent combination of abiotic and biotic factors. Typically, sweepstakes recruitment produces chaotic patchiness in allele distribution at small temporal and spatial scales (Avise 2000). While collections from single localities in multiple years were not attempted, the samples collected in this study span several years effort and (for Brazil and the southeastern U.S.) involved multiple localities. In contrast to chaotic patchiness, a consistency of pattern across years and samples were observed (e.g., the widespread abundance of haplotype A and the presence of the same haplotypes in South Carolina specimens collected in 1994 and Florida specimens collected in 1999). Population fluctuations also eliminate mtDNA lineages, such that the female effective population size may be orders of magnitude lower than indicated by census size. A survey of haplotypes in such a species may therefore reveal a large number of closely related haplotypes. If rock hind represents a single species with significant population structure (in mtDNA), then pairwise comparisons between haplotypes would create an L-curve of similar shape to the expected Poisson distribution. The additional smaller peaks in the mismatch distributions (an abundance of large interhaplotype distances; Figure 2-4) could be caused by rapid population expansion (Rogers et al. 1996). Overall, a
28 waveform result may indicate a rapid recovery from a recent genetic bottleneck (Rogers and Harpending 1992). Similar patterns have been observed in other tropical fishes, such as the mullet (Rocha-Olivares et al. 2000) and parrotfishes (Dudgeon et al. 2000). These data support the growing body of evidence for large-scale instability of marine populations over recent evolutionary time (e.g., Bowen and Grant 1997, Bucklin and Wiebe 1998, Grant and Bowen 1998, Lavery et al. 1996, Zane et al. 1998). The known species distribution of rock hind and greater soapfish implies a connection (either past or present) among reef habitats in multiple tropical biogeographic provinces and regions. There are considerable differences in the degree of genetic heterogeneity between species and between localities. Therefore an examination of the haplotype distribution in a geographic context may elucidate both potential differences in life history between these reef fishes as well as the effectiveness of biogeographic barriers. Biogeographic Barriers: Caribbean versus Brazil The reef-building corals of Brazil are highly distinct from those of the Caribbean (see Figure 49; Veron 1995), providing longstanding evidence for recognition of those areas as separate, large-scale biogeographic regions (Briggs 1974). Recent studies have documented that Caribbean and Brazilian reef provinces are also highly distinct for fish species, in terms of their distributions (Floeter and Gasparini 2000; Floeter et al. 2001; Rocha and Rosa 2001), taxonomic distinctions (Rocha and Rosa 1999), and evolutionary genetics (Bernardi et al. 2000, Bowen et al. 2001, Muss et al. 2001, Rocha et al. 2002). In contrast, comparisons in both species from western Atlantic localities did not reveal significant pairwise differentiations between the West Indian (i.e., not including the Florida/South Carolina rock hind) and Brazilian biogeographic provinces. However,
29 this may be influenced by the smaller sample sizes in the West Indian samples (with the important exception of 19 rock hind from Barbados). The sharing of haplotypes between the two provinces in both species studied here (Tables 2-2, 2-3) indicates recent or ongoing population connections. Are these species capable of routinely crossing the Amazonian plume vicariant barrier (described above)? One mechanism to accomplish this is that the extent of these speciesâ€™ PLD (presently unknown for both species) may allow recruitment of Brazilian haplotypes into the Caribbean (Figure 2-1). However, a prolonged PLD alone does not explain the presence of population genetic structure elsewhere in the range. Perhaps larval transport is not the only method of dispersal between Brazilian and Caribbean provinces. Benthic (juvenile and adult) stages could occupy sparse hard-bottom habitat between the Caribbean and northeast Brazil, as they do in subtropical waters of the eastern United States (Randall 1967). Alternately, they could occupy certain soft-bottom habitats such as the sponge communities under the Amazon and Orinoco plumes (Collette and Rtzler 1977). The possibility of dispersal (or population connections) through alternative habitats is especially relevant for soapfishes. Effective dispersal between the Caribbean and Brazilian reef systems requires survival across long distances without coral reef habitat, but also the capability to withstand or avoid the immense halocline imposed by the Amazon-Orinoco outflows (Longhurst 1998). The ability to exist in hyposaline waters is not unknown in the soapfishes, and is in fact a unifying character of Rypticus nigripinnus and R. bicolor in the Pacific (Guimares 1999; D. R. Robertson, STRI Panama, pers. obs.). Perhaps other soapfishes have wide salinity tolerances, as indicated by the
30 presence of Rypticus randalli in mangrove habitats in the Hawaiian archipelago (D. Greenfield, U. Hawaii, pers. comm.). Courtenay (1967) suggested that R. saponaceous might create shallow burrows in mud, although whether this actually occurs is not known (Guimares 1999). Nonetheless, the genetic findings for R. saponaceous is consistent with a Caribbean-Brazil link via softbottom habitat. Biogeographic Barriers: West versus East Atlantic The reef fishes of the East Atlantic have been understudied, and it is certain that additional species await discovery. In the comparisons of target species across the tropical Atlantic Ocean, the rock hind samples were characterized by the widespread distribution of haplotype A in Brazil, the mid-Atlantic Ridge, and East Atlantic. While this may be due to demographic processes unrelated to vicariance (see Discussion above), the nearly ubiquitous occurrence of haplotype A indicates that the rock hind has been able to colonize across long distances in the recent past. In contrast, the greater soapfish had moderate population structure (ST = 0.327 to 0.424, Table 2-5) between Brazil and the central Atlantic, and a strong (possibly species-level) distinction between the East Atlantic populations and elsewhere (pairwise STâ€™s > 0.90). The slight differentiations between haplotypes 1, 5, and 18 indicate a very rapid coalescence to an ancestor capable of colonization across the western-central Atlantic (Table 2-4). Collectively, these three haplotypes extend from Grenada and Barbados to Brazil to Ascension and St. Helena, and may support the â€˜stepping-stoneâ€™ colonization pattern across the open ocean barrier of the central Atlantic. St. Helena exhibited a very small haplotype diversity relative to other greater soapfish samples: 14 of 16 specimens had haplotype 18, implying either highly differential survival of lineages, or too little
31 time for divergence to occur in St. Helena soapfish in the cytochrome b region. Similarly, the lowest haplotype diversities in rock hind were in the Ascension and St. Helena samples. The low haplotype diversity in the Ascension/St. Helena province, coupled with the restricted haplotype distributions of soapfish on So Tom suggest that intermediate localities may be used as â€˜stepping-stones,â€™ but the lack of widespread haplotype in the greater soapfish indicates that this reef fish is not as able to cross the Atlantic as readily as the rock hind. The reason for the difference in oceanic dispersal ability between rock hind and greater soapfish is not readily apparent. Pelagic larval duration, while a tempting potential predictor of population structure, is one of many factors influencing genetic differentiation. The roles of oceanographic currents, freshwater plumes, recruitment success, and niche competition surely play a part in the connections and divisions among Atlantic coral reef fauna. Although each species undoubtedly meets unique challenges the lack of concordance might be enhanced by stochastic processes such as rare long distance dispersal, sweepstakes recruitment, local extirpation, and genetic drift in small founder populations. This is especially relevant to the rock hind, in which there is evidence for an evolutionary separation (approximately 3.2 MY) within the Caribbean Sea, but evidence of common ancestry between Brazil and West Indies populations. While this could be due to vicariance, other demographic and historical processes may also explain a pattern that appears to be very unusual among Caribbean reef fishes. The distribution of mtDNA diversity in both rock hind and greater soapfish offers insight into how biogeography influences speciation in the marine realm. The geography
32 and oceanography of the Atlantic basin may promote intraspecific (or intrageneric) partitions (Banford et al 1999, Graves 1995, Muss et al 2001, Tringali and Wilson 1993). The distances among these biogeographic provinces and the durations of pelagic larvae are not the only relevant factors in shaping patterns of genetic structure in panoceanic species. Stochastic colonizations and utilization of isolated islands allow for contact between Atlantic regions. The different responses to the mid-Atlantic and the Amazon-Orinoco barriers by two confamilial and even congeneric species (e.g., Rocha et al. 2002) illustrate how complex interactions may be involved in â€œsimpleâ€ predictors of population structure (e.g., dispersive larvae, adults restricted in habitat type, modern surface current patterns, straight-line distances between habitats). Rock hind and greater soapfish thus provide examples of the dynamic evolutionary processes of population isolation and differentiation which may typify the superficially continuous distributions of marine organisms.
Table 2-1. Genetic variation in the rock hind Epinephelus adscensionis and greater soapfish Rypticus saponaceous E. adscensionis R. saponaceous Location n h n h South Carolina 3 0.667 0.314 0.002 0.002 â€” â€” â€” Florida 9 0.889 0.091 0.003 0.002 â€” â€” â€” Bahamas 5 1.000 0.127 0.004 0.003 â€” â€” â€” Grenada â€” â€” â€” 5 0.900 0.161 0.003 0.002 Barbados 19 0.731 0.109 0.001 0.001 4 1.000 0.177 0.003 0.002 Brazil 16 0.625 0.139 0.002 0.002 23 0.913 0.035 0.003 0.002 Trindade 2 0.000 0.000 0.000 0.000 â€” â€” â€” Ascension 21 0.186 0.110 0.000 0.000 11 0.873 0.089 0.003 0.002 Saint Helena 17 0.228 0.130 0.001 0.001 16 0.242 0.135 0.001 0.001 Cape Verde â€” â€” â€” 18 0.863 0.042 0.004 0.002 So Tom 17 0.331 0.143 0.001 0.001 9 0.722 0.097 0.003 0.002 Total: 109 0.586 0.058 0.008 0.004 86 0.929 0.016 0.019 0.009 33 Note: The number of specimens (n), the haplotype diversity (h) and nucleotide diversity () for each locality and in total are listed. Some species were not collected (â€”) at some localities. Sample size at Trindade I. (n = 2) precludes statistical treatment.
Table 2-2. Haplotype occurrences and sample sizes observed in Epinephelus adscensionis South Carolina Florida Bahamas Barbados Brazil Trindade Ascension St. Helena So Tom Total Haplotype n = 3 9 5 19 16 2 21 17 17 109 A â€” â€” â€” 10 10 2 19 15 14 70 B 2 â€” â€” â€” â€” â€” â€” â€” â€” 2 C 1 â€” â€” â€” â€” â€” â€” â€” â€” 1 D â€” 3 â€” â€” â€” â€” â€” â€” â€” 3 E â€” 2 â€” â€” â€” â€” â€” â€” â€” 2 F â€” 1 â€” â€” â€” â€” â€” â€” â€” 1 G â€” 1 â€” â€” â€” â€” â€” â€” â€” 1 H â€” 1 â€” â€” â€” â€” â€” â€” â€” 1 I â€” 1 â€” â€” â€” â€” â€” â€” â€” 1 J â€” â€” 1 â€” â€” â€” â€” â€” â€” 1 K â€” â€” 1 â€” â€” â€” â€” â€” â€” 1 L â€” â€” 1 â€” â€” â€” â€” â€” â€” 1 M â€” â€” 1 â€” â€” â€” â€” â€” â€” 1 N â€” â€” 1 â€” â€” â€” â€” â€” â€” 1 O â€” â€” â€” 2 â€” â€” â€” â€” â€” 2 P â€” â€” â€” 1 â€” â€” â€” â€” â€” 1 Q â€” â€” â€” 1 â€” â€” â€” â€” â€” 1 R â€” â€” â€” 1 â€” â€” â€” â€” â€” 1 S â€” â€” â€” 1 â€” â€” â€” â€” â€” 1 T â€” â€” â€” 1 â€” â€” â€” â€” â€” 1 U â€” â€” â€” 1 â€” â€” â€” â€” â€” 1 V â€” â€” â€” 1 â€” â€” â€” 1 â€” 2 W â€” â€” â€” â€” 1 â€” â€” â€” â€” 1 X â€” â€” â€” â€” 1 â€” â€” â€” â€” 1 Y â€” â€” â€” â€” 1 â€” â€” â€” â€” 1 Z â€” â€” â€” â€” 1 â€” â€” â€” â€” 1 AA â€” â€” â€” â€” 1 â€” â€” â€” â€” 1 34
Table 2-2. Continued South Carolina Florida Bahamas Barbados Brazil Trindade Ascension St. Helena So Tom Total Haplotype n = 3 9 5 19 16 2 21 17 17 109 BB â€” â€” â€” â€” 1 â€” â€” â€” â€” 1 CC â€” â€” â€” â€” â€” â€” 1 â€” â€” 1 DD â€” â€” â€” â€” â€” â€” 1 â€” â€” 1 EE â€” â€” â€” â€” â€” â€” â€” 1 â€” 1 FF â€” â€” â€” â€” â€” â€” â€” â€” 1 1 GG â€” â€” â€” â€” â€” â€” â€” â€” 1 1 HH â€” â€” â€” â€” â€” â€” â€” â€” 1 1 35
36 Table 2-3. Population pairwise differentiation observed in Epinephelus adscensionis Florida Bahamas Barbados Brazil Ascension St. Helena So Tom Florida + + + + + Bahamas 0.911 + + Barbados 0.947 0.165 Brazil 0.937 0.125 0.039 Ascension 0.973 0.363 0.043 0.055 St. Helena 0.962 0.238 0.018 0.025 0.007 So Tom 0.962 0.246 0.028 0.030 0.007 0.000 Note: The pairwise ST values are given below the diagonal, while above the diagonal are significant (+) or not significant () pairwise differentiation as observed under Markov chain simulation, based on an overall P < 0.05 after full Bonferroni correction. Trindade and South Carolina samples were not included because of low sample sizes.
37 Table 2-4. Absolute haplotype frequencies and sample sizes in Rypticus saponaceous Grenada Barbados Brazil Ascension St. Helena Cape Verde So Tom Total Haplotype n = 5 4 23 11 16 18 9 86 1 2 1 4 â€” â€” â€” â€” 7 2 1 â€” â€” â€” â€” â€” â€” 1 3 1 â€” â€” â€” â€” â€” â€” 1 4 1 â€” â€” â€” â€” â€” â€” 1 5 â€” 1 3 â€” 1 â€” â€” 5 6 â€” 1 â€” â€” â€” â€” â€” 1 7 â€” 1 â€” â€” â€” â€” â€” 1 8 â€” â€” 5 â€” â€” â€” â€” 5 9 â€” â€” 3 â€” â€” â€” â€” 3 10 â€” â€” 1 â€” â€” â€” â€” 1 11 â€” â€” 1 â€” â€” â€” â€” 1 12 â€” â€” 1 â€” â€” â€” â€” 1 13 â€” â€” 1 â€” â€” â€” â€” 1 14 â€” â€” 1 â€” â€” â€” â€” 1 15 â€” â€” 1 â€” â€” â€” â€” 1 16 â€” â€” 1 â€” â€” â€” â€” 1 17 â€” â€” 1 â€” â€” â€” â€” 1 18 â€” â€” â€” 4 14 â€” â€” 18 19 â€” â€” â€” 2 â€” â€” â€” 2 20 â€” â€” â€” 1 â€” â€” â€” 1 21 â€” â€” â€” 1 â€” â€” â€” 1 22 â€” â€” â€” 1 â€” â€” â€” 1 23 â€” â€” â€” 1 â€” â€” â€” 1 24 â€” â€” â€” 1 â€” â€” â€” 1 25 â€” â€” â€” â€” 1 â€” â€” 1 26 â€” â€” â€” â€” â€” 5 2 7 27 â€” â€” â€” â€” â€” 3 4 7 28 â€” â€” â€” â€” â€” 3 3 6 29 â€” â€” â€” â€” â€” 3 â€” 3 30 â€” â€” â€” â€” â€” 2 â€” 2 31 â€” â€” â€” â€” â€” 2 â€” 2
38 Table 2. Population pairwise differentiation observed in Rypticus saponaceous Grenada Brazil Ascension St. Helena Cape Verde So Tom Grenada + + Brazil 0.128 + + + + Ascension 0.456 0.327 + + St. Helena 0.695 0.424 0.005 + + Cape Verde 0.906 0.909 0.911 0.938 So Tom 0.929 0.919 0.928 0.963 0.042 Note: The pairwise ST values are given below the diagonal, while above the diagonal are significant (+) or not significant () pairwise differentiation as observed under Markov chain simulation, based on an overall P < 0.05 after full Bonferroni correction. Barbados specimens were not included because of low sample size.
39 Figure 2-1. Biogeographic divisions of the tropical Atlantic with sample localities. Collection localities of greater soapfish Rypticus saponaceous (shaded triangles), rock hind Epinephelus adscensionis (open triangles), or both species (circles). The tropical biogeographic divisions discussed in the text are indicated as follows: the West Indian Province including Bermuda (gray stippled area); the Caribbean Province (western solid line); the Brazilian Province (gray line); the Ascension-St. Helena Province (double circles); and the East Atlantic Region including Cape Verde (eastern solid line and circle). The dotted gray line indicates the area of intrusion by Amazonian discharge, the Brazilian province north of this area is devoid of coral-type reefs.
40 Figure 2-2. Distances among Epinephelus adscensionis haplotypes (Aâ€“HH) and selected congeners, as indicated by a neighbor-joining (NJ) tree. For those nodes with > 70% retention after 500 bootstrap iterations, the bootstrap support is indicated at the node.
41 Figure 2-3. Distances among Rypticus haplotypes (1) and selected congeners, as indicated by a neighbor-joining (NJ) tree. For those nodes with > 70% retention after 500 bootstrap iterations, the bootstrap support is indicated above the node.
42 Figure 2-4. Frequency distributions for pairwise haplotype distances for cytochrome b from Epinephelus adscensionis and Rypticus saponaceous.
CHAPTER 3 PHYLOGENETIC RELATIONSHIPS AMONG SOME GROUPER AND CORALGROUPER SPECIES, WITH COMMENTS ON THE TRIBE EPINEPHELINI (PERCOIDEI: SERRANIDAE: EPINEPHELINAE). Introduction The tribe Epinephelini (the groupers, coralgroupers, coneys, and lyretails) is a set of tropical and subtropical marine fishes easily recognized by two confluent dorsal fins and an oblong body adapted to demersal predation (Cavalcanti 1999). Hermaphroditism, group spawning, relatively slow maturation, and low fecundity are common to many grouper species (Domeier and Colin 1997), making these apex predators particularly vulnerable to overfishing (Musick et al. 2000, Sadovy, 2001). Several smaller species are popular as aquarium fish, while the excellent food quality of the larger groupers supports a variety of fisheries, from cyanide fishing in the Indo-Pacific (IMA Indonesia 2001) to extensive line and trap fisheries in Europe and the Americas (Coleman et al. 2001). The marine commercial catch of groupers for the United States alone has a mean annual worth of nearly $3 million, not including the equal or greater recreational catch (NMFS 2002). The management of multispecies reef fish fisheries, the identification of biodiversity â€˜hotspotsâ€™ for conservation, and studies of comparative biology all require a taxonomy that reflects biological history (Collette and Vecchione 1995, Sweijd et al. 2000, Vecchione et al. 2000), but currently no such phylogenetic resolution is available. Groupers and their allies are members of the percoid family Serranidae, possibly a nonmonophyletic family and "classificatory waste-basket" (Johnson 1983). The limits of this family were established by Gosline (1966), who diagnosed three serranid 43
44 subfamilies, the Anthiinae, Serraninae, and the grouper subfamily Epinephelinae. Kendallâ€™s (1976) examination of predorsal bones and associated pterygiophores identified two evolutionary trajectories within Serranidae: an ancestral lineage including subfamilies Serraninae and Anthiinae and a derived lineage including the Epinephelinae with the Grammistinae, the latter uniting the soapfishes and their allies by combination of the Liopropominae, Grammistidae, and Pseudogrammidae of Gosline (1960). The subfamily Epinephelinae was presumed monophyletic because of the absence of an autogenous distal radial on the first dorsal-fin pterygiophore (Johnson 1983). Further modification was suggested by Johnson (1988), who included the soapfish tribes Diploprionini, Liopropomini, and Grammistini as derived taxa within Epinephelinae, and removed the percichthyid species Niphon spinosus into a fifth epinepheline subfamily, the Niphonini. The Niphonini was proposed as the common ancestor to the soapfish tribes and the monophyletic epinephelins. Please note that â€˜epinephelinesâ€™ refers to the grouper-soapfish subfamily Epinephelinae (following the usage of Johnson 1988), while â€˜epinephelinâ€™ denotes the grouper tribe Epinephelini. Support for a monophyletic Epinephelini was proposed as the shared presence of elongate spinelets on the second dorsal and caudal fins in postflexion larvae (Johnson and Keener 1984). Baldwin and Johnson (1993) investigated soapfish tribes and genera, but did not investigate the epinephelins because of a prohibitive lack of material. The literature review by Heemstra and Randall (1993) elevated the tribe Epinephelini to a monophyletic subfamily (Table 3-1). No phylogenetic analysis was presented, but characters supporting a monophyletic Epinephelinae included possession of: 10 precaudal and 14 caudal vertebrae, one supernumeray spine on the first dorsal fin pterygiophore, well-developed supramaxilla,
45 depressible inner teeth, a membrane attaching the proximal third of inner pelvic fin ray to the body, and larvae with the second dorsal fin spine and pelvic fin spine elongated and serrate (Heemstra and Randall 1993). These features define the ingroup for consideration in this study and are treated as members of the serranid tribe Epinephelini (sensu Johnson 1983). Major difficulties in elucidating epinephelin relationships are the tribeâ€™s circumtropical distribution, impressive diversity (159 species in 15 genera sensu Heemstra and Randall 1993) and, like most coral reef fishes, a near total lack of fossil record (Bellwood and Wainwright 2002). The Epinephelini includes several complexes of sympatric and parapatric species distinguished primarily by color pattern and combinations of overlapping meristic characters (e.g., the reticulated groupers of the Indo-West Pacific). Some species pairs differ only in distribution and color pattern (e.g., Epinephelus multinotatus, E. cyanopodus, E. trophis, and E. flavocaeruleus), while marked color pattern or meristic differences exist between distant populations within single widely distributed species (e.g., Epinephelus fasciatus, Mycteroperca acutirostris). Recognized species may consist of separate noninterbreeding units (Carlin et al. in press, Gilles et al. 2000, Heemstra and Randall 1993) and hybrids have been described between some species and genera (Bostrom et al. 2000, Glamuzina et al. 2001). Resolution of such species complexes by molecular techniques addresses not only evolutionary history (e.g., McMillan and Palumbi, 1997) but allows for more informed management in multispecies fisheries (Collette and Vecchione 1995, Sweijd et al. 2000). The number of accepted genera differs between Atlantic and Indo-Pacific biologists (e.g., Grove and Lavenberg 1997, Humann et al. 2002, Robins et al. 1986, Sadovy and
46 Cornish 2000). Smith (1971) used phenetic analysis of meristic and skeletal features to arrange American grouper species into a phylogenetic hypothesis that proposed early separate divergences of Paranthias and Mycteroperca from an Epinephelus lineage. The Epinephelus lineage then diverged nearly simultaneously among Cephalopholis, Alphestes, Dermatolepis, and a lineage that eventually gave rise to all American Epinephelus and the genus Promicrops. Based upon this hypothesis, the four genera derived from the Epinephelus lineage were demoted to subgenera. Examinations of spinelet development in larvae upheld the subgeneric demotions of Smith (1971), but with a few exceptions, especially within Cephalopholis and Alphestes (Johnson and Keener 1984). Inclusion of the Pacific genus Plectropomus rooted a cladogram using 12 larval characters, contradicting an early divergence of Mycteroperca and Paranthias (Leis 1986). Instead, Plectropomus was the most ancestral genus with available larvae, and a trichotomy of Mycteroperca, Paranthias, and Epinephelus was the most derived, with Gonioplectrus (based upon an incomplete series) and Cephalopholis as an intermediate genus (Leis 1986). The binomial nomenclature of Heemstra and Randall (1993) was chosen as a reference taxonomy for this study. Systematic resolution of the epinephelines and related taxa would do more than standardize a long-standing taxonomic debate. Comparative studies of fisheries ecology, reproductive biology, behavior, or functional morphology with a phylogenetic component require knowledge of sister taxon relationships to determine character polarity. The epinepheline fishes usually exhibit large size, aggressive predation, a tendency to form spawning aggregations, relatively slow growth, and low fecundity; all of which make epinephelines both ecologically important as primary demersal predators and highly
47 vulnerable to overexploitation (Coleman et al. 2001, Musick et al., 2000). Fisheries regulations are, in part, based on knowledge assuming that field identifications are accurate and that recognized species are real biological entities. Such assumptions are also made in attempts to delineate areas of endemicity, speciation, and biodiversity loss (Gill and Kemp 2002, Roberts et al. 2002). Recently, the veracity of these assumptions in large marine species complexes has been questioned, with the six races of Epinephelus fasciatus (sensu Heemstra and Randall 1993) suggested as a prominent example of a cryptic complex waiting to be discovered (Gill and Kemp 2002). In this study, I tested the validity of tribe Epinephelini by comparing posterior probability (Bayesian) and maximum parsimony (MP) gene trees from a partial sequence of the mitochondrial cyt b gene and intron 6 of the nuclear lactose dehydrogenase gene. The phylogenetic status of species was evaluated by replicate sampling within a priori species whenever possible. Specifically, I generated gene trees to: evaluate the utility of cyt b and LDHA6 gene trees as markers for the subfamily level in percoids; test the accepted monophyly of the grouper and coralgrouper tribe Epinephelini; determine the most likely phylogenetic position of the epinephelin genera Aethaloperca, Cromileptes, Saloptia, and Variola (which have no genetic data in the literature); and test if the molecular data observed here are as likely to be explained by other published phylogenetic hypotheses (Figure 3-1; Craig et al. 2001, Leis 1986, Smith 1971). I then suggest systematic revisions where appropriate and present a phylogenetic hypothesis for further testing. Materials and Methods Tissue samples were obtained from multiple localities per species whenever possible, collected from port samples or during snorkel and SCUBA diving using
48 polespears and microspears. The Ichthyology Division of the Kansas University Museum of Natural History (KU) provided some tissue samples. In most locations, gill and/or muscle samples were taken and stored in a saturated-salt DMSO buffer (Amos and Hoelzel 1991). Several samples each were chosen to represent the range of both geographic distance and genetic divergence observed across the species range of Epinephelus adscensionis and Rypticus saponaceous (Carlin et al. in press). Total genomic DNA was isolated either by organic extraction (phenol:chloroform: isoamyl alcohol) and 95% ethanol precipitation, or by using the DNeasy Isolation Kit (Promega, Inc., Madison WI). Isolated DNA was resuspended in 200 mL TE (10 mM Tris and 1 mM EDTA, pH 8.0). An approximately 750-base pair (bp) fragment of the mitochondrial cyt b gene was amplified via the polymerase chain reaction (PCR; Saiki et al. 1985) with primers Cyb-09H (5'-GTGACTTGAAAAACCACCGTTG-3'; Song 1994) and Cyb-07L (5'-AATAGGAAGTATCATTCGGGTTTGATG-3'; Taberlet et al. 1992). The amplification usually included 3 mM MgCl, 40 nM of each primer, 17.5 nM of each dNTP, 2.30 units of Taq DNA polymerase (Promega, Inc.), and 3.5 L of 10x PCR buffer (Promega, Inc.) in 35 L total volume. Amplification temperature regime included an initial 3-minute denaturation at 94C and final extension of 72C for 3-minutes, with 25 cycles of 30-seconds at 94C, 40-seconds at 48C, and 1-minute at 72C. The nuclear LDHA6 intron was directly amplified from genomic DNA isolations using primers LDHA6F and LDHA6R, with amplification conditions varying only slightly from published protocol (Quattro and Jones 1999). Oligonucleotide primers were removed by Ultrafree-MC centrifugal filter units (Millipore Corp., Billerica MA) and/or simultaneous digestion with exonuclease I and
49 shrimp alkaline phosphatase (USB Corp., Cleveland OH; Silva et al. 2001). Amplification products were denatured and DNA sequencing reactions were conducted with fluorescently labeled dideoxy terminators according to manufacturer's recommendations (ABI Model 800, Applied Biosystems, Inc., Foster City, CA). Labeled extension products were gel separated and analyzed with an automated sequencer (ABI Models 310, 373A and 377) by the author, the University of Florida Sequencing Core, and/or the Nevada Genomics Center. All samples were sequenced at least twice with amplification primers; some cyt b samples were sequenced internal to the amplification targets using primers Cyb-19H (5'-CTCACTGCTCGGACTCTG-3'), Cyb-21L (5â€™-GTTTTGATGTGTGGTGGAG T-3â€™), or Cyb-22H (5'-GTGAARTTGTCTGGGTCT CCT-3'). Resulting chromatograms were edited using Sequencher ver. 3.0 (Gene Codes Corp., Ann Arbor, MI) and combined with cyt b sequences in GenBank from perciform fishes. Sequences were aligned by sum-of-pairs progressive pairwise alignment using the Needleman-Wunsch (1970) algorithm in Sequencher and manually inspected. Intron chromatograms were visually examined for multiple peaks at a single nucleotide position. Such samples were presumed as representing heterozygotes and removed from further analyses. Presumed homozygote (single signal) samples were sequenced at least twice. Sequences were compared to those in GenBank by BLAST alignment. GenBank accessions of serranids, as well as any percoid accessions with 85% sequence match by Sequencher included in the final alignment. Nucleotide polymorphism, pairwise distances between sequences, and phylogenetic estimations were performed using Paup*5.10 (Swofford 2002). A 2 test of base
50 frequency homogeneity was performed considering all positions and third nucleotide position only, with the acknowledgment that this test assumes phylogenetic independence. Phylogenetic signal was examined by testing for signal among random trees (g1 test; Hillis et al., 1996) and visually assessed by plotting transitions against transversions and proportional distance (p-distance). Three partition homogeneity tests with 100 replicates each (Farris et. al. 1994, 1995) were used to test for incongruence between the nuclear intron and mitochondrial cytochrome b loci. The tests considered: epinephelin taxa with sequence data at both loci; all taxa with data at both loci; and all available taxa, including those lacking sequenced data for one locus. Single-locus alignments include comparisons at multiple taxonomic levels, ensuring different amounts of relevant signal for different questions. Therefore, signal assessment was repeated with increasingly more inclusive sets of taxa, considering all data versus alternate codon positions (for cyt b) and indel states (LDHA6). Additionally, numbers of variable, parsimony-informative, and parsimony-uninformative sites were calculated for all taxa, the subfamily Epinephelinae only, and tribe Epinephelini only. Neighbor joining (NJ) trees were constructed using proportional distances (p-distances) and then resampled by 500 bootstraps. Sequences united at a p-distance NJ bootstrap frequency >90% were assumed to be natural groupings and to maximize computational efficiency, nodes very highly supported under a simple NJ model subsequently defined topological constraints imposed on all parsimony analyses. Two separate methods of phylogenetic estimation were used: maximum parsimony and posterior probability analysis. A partial search of available treespace was conducted under the maximum parsimony criterion, following the recommendations of Takahashi
51 and Nei (2000), for the given sample size and variance in internode length. Cladograms were constructed by 50 iterations of heuristic searches by random taxon addition from a random tree with the NJ constraints (see above), with each additional iteration followed by the tree bisection-reconnection branch swapping protocol (Swofford and Begle 1993). Strict and majority-rule consensus trees were constructed from equally parsimonious MP trees. Three arbitrarily chosen weighting schemes were utilized in analyses of cyt b: considering all sites with equal weights, weights reflecting the transition:transversion bias previously reported in fish cyt b studies (Song 1994), and weighting incorporating both a codon position bias (3:5:1 for first, second and third positions) and a 1:4 ti:tv bias. The LDHA6 locus was analyzed both with indels (equal weights at all sites, 1:2 ti:tv, 1:2 ti:tv and 1:5 nucleotide:indel) and without indels (equal weights at all sites, 1:2 ti:tv). Posterior probability (Bayesian) analysis was conducted by the Monte Carlo Markov chain algorithm in MrBayes 2.01, which approximates the results of a likelihood analysis with bootstrap resampling (Huelsenbeck and Ronquist 2001). Two separate Bayesian analyses evolved four chains (three heated, one cold) from an initial random tree for 2 million generations, sampling a tree every 200 generations. The first Bayesian analysis ran with three chains heated at power=0.2 with estimated priors. The second analysis involved three chains heated at power=0.4 with unknown (flat) priors (e.g., equal rates of substitution across sites). The first 10% of generated trees were discarded as burn-in, and consensus phylograms (strict and majority-rule) were constructed from the remaining trees. The prior probabilities for some Bayesian analysis (power=0.2) were generated from a model of DNA substitution for each locus. Model selection utilized Paup*5.10
52 (Swofford 2002) and Modeltest 2.0 (Posada and Crandall 1998). A bottom-up hierarchical likelihood ratio test (LRT; Cunningham et al. 1998, Posada 2001) selected substitution models at each level of comparison. The LRT was conducted among epinephelin sequences only with one randomly chosen sequence per a priori species to minimize model estimation error from incomplete taxon sampling. Topological comparisons were carried out in two steps. First, gene trees generated under different weighting schemes were compared at each locus to best estimate sequence evolution, using the Shimodaira-Hasegawa (SH) test (Shimodaira and Hasegawa 1999). The SH test corrects for multiple comparisons among a priori trees generated from the same dataset but may be overly conservative, with a probability of correct rejection being less than (Shimodaira 2002, Shimodaira and Hasegawa 1999). The tree construction method indicated as optimal by the SH test was used to generate new gene trees but with additional topological constraints as suggested by a priori hypotheses of relationship (e.g., Craig et al. 2001, Johnson 1983). A comparison of the tree with minimal constraints versus the a priori topology was performed using the Kishino-Hasegawa (KH; Kishino and Hasegawa 1989) test. The SH and KH tests were performed in Paup*v.5.10; both indicate a topology as optimal if significant improvement in likelihood score is found relative to an alternate topology. Results A total of 52 identified epinepheline species were available from 11 genera (out of a possible 159 species in 15 genera listed by Heemstra and Randall (1993). Three epinepheline genera (comprised of 7 described species) were not available for analysis. Potential outgroup taxa included the subfamily Serraninae (9 species from 3 genera) and
53 the subfamily Grammistini (2 species from 2 genera). BLAST alignments of GenBank sequences from 10 percoid families resulted in successful alignments among generated sequences and accessions in the perciform families Moronidae, Percichthyidae and Polyprionidae. The cyt b alignment also included six haplotypes identified only to genus: Epinephelus GA-1999 (GenBank Accession AF143193; Allegrucci et al. 1999) and Epinephelus sp. A (collected by the author from commercial markets in Hong Kong), 1 Mycteroperca sp. A (mid-Atlantic islands of the Azores, Portugal), 2 specimens of Serranus sp. A (Gulf of Guinea, East Atlantic; from the tissue collection of D. R. Robertson, STRI-Panama), and 1 Aporops spp. (KU 804; an unidentified specimen of 6 mm TL). All sequence data were combined into a single alignment and tested for incongruence by partition homogeneity testing. A significant incongruence (P = 0.010) was detected between the nuclear and mitochondrial data when testing all taxa (n = 87), all taxa sequenced at both loci (n = 71), and epinephelin taxa with sequence data at both loci (n = 69). Subsequent results are from separate analyses of cytochrome b and LDHA6. Cytochrome b A total of 118 individuals yielded 87 unique haplotypes for 665 bp from the cyt b mitochondrial gene. The alignment was poor in guanines (Table 3-2), similar to studies of cyt b in other percoid families (e.g., Near et al. 2000, Song et al. 1998). DNA substitution frequencies varied by steric type (transition vs. transversion) and codon position, indicating multiple substitutions per site (i.e., saturation), with all or nearly all third position sites parsimony informative (Table 3-2). The number of transitions (ti) did not increase linearly with the transversions (tv) in plots of absolute numbers of
54 substitutions for most pairwise comparisons. Instead, pairwise comparisons of all taxa while considering all codon positions have a wide range of transversions (from 1 to 78) if more than 45 transitions are present between compared taxa. All nucleotide sites in the third codon position were variable. The g1 tests detected significant phylogenetic signal for all comparisons (P = 0.010). The LRT resulted in selection of a general time reversible (GTR) model (Hillis et al. 1996) incorporating a correction ( = 0.584) and some sites (47.05%) invariable. Two or more specimens were sequenced for 22 species (42.30% of species surveyed). Mean pairwise divergence between interspecific haplotypes within genera was d = 13.90 % (range: 0.02 to 18.9%). The greatest differentiation within a genus occurred between haplotypes of Epinephelus adscensionis and E. itajara (a difference of 35 ti and 90 tv). The least interhaplotype divergence estimates occurred between Plectropomus leopardus vs. P. maculatus (9 ti, 0 tv), Epinephelus niveatus vs. E. niphobles (10 ti, 0 tv), Mycteroperca bonaci vs. M. rosacea (8 ti, 3 tv), and E. chlorostigma vs. E. cyanopodus (47 ti, 4 tv). Multiple haplotypes within a single species were monophyletic for each species (>98% support after 500 bootstrap and 500 jackknife iterations). However, strong support was also recorded for deep divergences within monophyletic Epinephelus adscensionis (d = 4.3%), Rypticus saponaceous (d = 4.4%), and Variola louti (d = 10.4%). Of the 24 species with >1 specimen, haplotypes were shared across oceans in Cephalopholis cruentata (Bahamas and So Tom), Epinephelus adscensionis (Bahamas and So Tom), Rypticus saponaceous (So Tom and Cape Verde; Ascension and Florida), and Variola louti (Australia and Seychelles; Australia and
55 Philippines). Two specimens of M. phenax from the Gulf of Mexico shared a haplotype with the unidentified M. spp. A from the Azores. Translation of the 87 haplotypes reduced the dataset into 66 unique amino acid sequences with 40 parsimony informative characters out of a possible 221 amino acids. A p-distance NJ tree resampled by 500 bootstrap iterations united two haplotypes of E. itajara, all Plectropomus sequences, and unified serranines with anthiines. The serranines, anthiines, and the outgroups Morone and Dicentrarchus were separated with poor (62%) bootstrap support from all epinepheline and grammistine amino acid sequences. Comparisons of cyt b amino acids may be able to distinguish percoid families and subfamilies, but does not have sufficient variation for more exclusive phylogenetic studies. Nodes supported in the amino acid alignment also had >97% support in the nucleotide p-distance NJ tree after 500 bootstrap iterations; such nodes were used to define topological constraints for parsimony analyses. MtDNA constraint nodes united all haplotypes within Alphestes, Hypoplectrus, Lateolabrax, Morone, Mycteroperca, Plectropomus, Variola, and all sequences within a priori species (as defined by Heemstra and Randall 1993) except for . Cyt b gene trees constructed under a variety of methods varied little in topology from the p-distance tree. Because MP cladograms were constructed using a constraint defined by NJ bootstrapping, some agreements between NJ and MP trees are biased. However no such constraints were defined for Bayesian analyses. Posterior probability explanations of the cyt b data provided the most likely explanations of the data (Table 3-3), although this tree was not significantly different by SH tests from either the p-distance phylogram or a MP cladogram using extensive weighting. Because SH tests could detect
56 few significant topological differences among alternate methods for this dataset, the Bayesian gene tree (power = 0.4) is presented (Figures 3-2 and 3-3), as this may provide a better estimate of branch lengths among taxa (Braun and Kimball 2002). Nuclear Intron Direct sequencing of LDHA6 amplicons resulted in unambiguous sequences for nearly a third of samples utilized in the cyt b analysis. The only nonepinepheline taxa to amplify at the LDHA6 locus were Rypticus saponaceous (tribe Grammistini) and Stereolepis gigas (family Polyprionidae). The lack of serranine, anthiine, and other outgroup amplicons may be due to mispriming; no combination of published LDHA6 primers nor custom primers designed from epinephelines resulted in amplification. Two amplification products, approximately 130 or 190 bp in length, were sequenced from 81 specimens from 52 species in the cyt b dataset. A total of 71 unique sequences were aligned manually under a bias against novel gaps, resulting in 67 of 186 positions occupied by indels for one or more sequence. The alignment had 70 indels and a nearly equal number of transitions and transversions (Table 3-2). Phylogenetic signal was assessed for the entire alignment (with indels as a fifth base in Paup*), the alignment without indels, and with an alignment of indel positions alone in a nucleotide presence/absence matrix. The greatest percentage of parsimony informative sites (94.7%) occurred in the indels-only matrix (Table 3-2). The nucleotides-only alignment was best explained in LRT by a GTR+I+ model with 39.3% invariant sites and a shape parameter = 0.870. Thirteen of the 22 species sampled with more than one specimen possessed >1 genotype per species. No genotypes were found in >1 species. Only 2 species (M. fusca
57 and R. saponaceous) of the 13 species variable at this locus were united strongly in a monophyletic group with strong support in a p-distance NJ tree after 500 bootstrap iterations. An indels-only NJ tree clustered alleles into three phylogroups. One group of 21 alleles included the percichthyid Stereolepis gigas, the epinepheline Aethaloperca rogaa, 4 Cephalopholis spp., and all Plectropomus specimens. The group of 40 alleles included Anyperodon, Cromileptes, and all Mycteroperca specimens. The Mycteroperca group has eight insertions 1 to 6-bp long and one 13-bp deletion relative to the Cephalopholis-Plectropomus group. A third group containing 5 Epinephelus spp. had an indel pattern intermediate to the Mycteroperca and Cephalopholis-Plectropomus groups, but with a unique 5-bp deletion. Almost no significant differences could be detected among gene trees for this locus (Table 3-4), and this is interpreted here as an indicator of the poor quality of phylogenetic information â€“ some nodes were supported at all trees and others collapsed into polytomies at all trees. The simplest model that incorporates all information available, a parsimony cladogram with equal weighting but considering both sequence and indel position data, is presented in Figure 3-4. Discussion Presented here is the largest collection of serranid DNA sequences to date, and substantial evidence for a nonmonophyletic Epinephelini (Figure 3-3). Robust gene trees are best constructed using a large amount of sequence data from as wide a geographic and taxonomic sampling as possible (Pollock et al. 2002, Wakeley 2000, Zwickl and Hillis 2002), but the resultant gene tree may still not reflect the true phylogeny (Nichols 2001, Pamilo and Nei 1988). The nearly 200 species necessary to completely resolve relationships among the groupers and their potential outgroups is prohibitive to most single investigations. The large number of sequences analyzed, and the widely varying
58 branch lengths due to the inclusion of both family and intraspecific comparisons dramatically lowered estimates of phylogenetic signal. However, the topologies of cyt b consensus trees constructed under alternate optimality criteria and different character weighting options varied only slightly, usually in the placement of the same highly divergent sequences (e.g., the monotypic genus Aethaloperca). Moreover, many of the relationships suggested by Craig et al. (2001) (e.g., polyphyly of the genus Epinephelus) are confirmed by this study. Significant incongruence exists between the intron alignment and the mitochondrial locus. Combining alignments allows examination of all available evidence (Wiens and Reeder 1995) and the ability of the partition homogeneity test to correctly identify incongruence has been questioned (Dolphin et al. 2000; Yoder et al. 2001). However, LDHA6 differs from cytochrome b by the having less than half the number of species present in the mitochondrial alignment. More problematical is proper alignment of the numerous gaps in the LDHA6 alignment. Two different size classes of LDHA6 allele were amplified. The longer alleles tended to form unresolved polytomies including both derived and ancestral mtDNA sequences. Some species shown as monophyletic in mtDNA haplotype with high bootstrap support (>95%) were not monophyletic at LDHA6. Variola louti specimens yielded either long or short LDHA6 alleles. The monophyly of each of these species is therefore challenged at this intron, with numerous undescribed polyphylys and paraphylys. Alternately, the length differences among alleles precludes accurate alignment, especially in placement of small gaps. Therefore separate analyses of cyt b and LDHA6 were conducted and subsequent phylogenies were compared to both eachother and to prior hypotheses of relationship.
59 Significant length differences were detected among all trees (Table 3-4), indicating a failure for either locus to support any hypothesis from alternate evidence. In contrast, visual comparison confirms that some relationships suggested at mtDNA loci (either cyt b or 16S) were not refuted by intron data, many relationships were supported by all three datasets, and many of the clades identified may have support in morphological characters as well. Concordance among alternate trees is illustrated in a majority-rule consensus tree (Figure 3-5) constructed from taxa common to the cyt b, LDHA6, and 16S (Craig et al. 2001) gene trees. Therefore the phylogeny in Figure 3-5 represents current understanding of grouper molecular evolution and provides a framework for examining intrafamily relationships. Evolution among tribes in the Serranidae has been summarized as a split between two evolutionary lineages, an ancestral anthiin-serranin lineage and a lineage from which first the Niphonini, then a monophyletic Epinephelini, then a monophyletic Grammistini were derived ( Johnson 1983, Johnson and Keener 1984, Kendall 1976). These data present robust support for a common grammistin-epinephelin lineage within the Serranidae (as suggested by Craig et al. 2001, Johnson 1983). Several species, sequenced directly and accessed from GenBank, were used to sample the anthiin-serranin lineage and basal percoid families. The Hypoplectrus, Paralabrax, and Serranus sequences formed a well-supported monophyletic group, albeit including a polyphyletic Serranus (the monophyly of Serranus may not be supported at other loci as well, M. McCartney, University of North Carolina at Wilmington, pers. comm.). Both loci support the ancestral position of Plectropomus and Saloptia relative to Epinephelus and its allies (Heemstra and Randall 1993, Leis 1986, Smith 1964).
60 Monophyly of the Epinephelini Viewed conservatively, the cyt b gene tree places the grammistins and the anthiin-serranin lineage into an unresolved trichotomy with the monophyletic epinephelin genus Variola. Together the three clades serve as a sister group to a lineage including both percichthyids and the rest of the epinephelin genera (Figure 3-3). Thus the placement of Plectropomus and Variola makes Epinephelini paraphyletic as defined by Johnson (1983). Three of the four Variola specimens in the LDHA6 gene tree (Figure 3-5) were sister taxa to alleles from the grammistin Rypticus and one allele from Plectropomus maculatus, termed an ancestral genus on the basis of larval morphology (Leis 1986), giving slight evidence for inclusion of Variola within the epinephelin-grammistin lineage as opposed to the anthiin-serranin lineage. The 16S gene tree by Craig et al. (2001) placed two species of Plectropomus as a sister group to the grammistins, making the tribe Epinephelini paraphyletic if Plectropomus was included. Plectropomus and Rypticus differ by approximately 12% at 16S (authorâ€™s estimate). Craig et al. (2001) upheld a monophyletic Epinephelini based upon morphological evidence of Johnson (1983), invoking long branch attraction to explain the Plectropomus-Rypticus affinity. The level of divergence separating the basal genera Plectropomus and Variola (Epinephelini) from Rypticus and Aporops (Grammistini) is at or close to saturation for both LDHA6 and cyt b (pairwise sequence divergences ranging from d = 25.2 to 26.6% and 16.5 to 20.2%, respectively). Long-branch attraction is to be expected at such great divergences, yet the grammistin sequences consistently clustered with Plectropomus (as opposed to Serranus or Stereolepis or Lateolabrax) at cyt b under different weighting schemes, including all bases, and not considering bases from the third codon position.
61 A similar clustering at a nuclear intron may be less coincidental. Johnson (1983) chose to place the ancestral Niphon spinosus into a separate tribe rather than as a basal taxon within Epinephelini. Such an arrangement correctly reflects the phylogenetic relationships if the Epinephelini is monophyletic. If the taxonomic classification is to continue to reflect genealogic patterns, then the genera Plectropomus, Saloptia, and Variola would be removed to a separate tribe, named after its oldest generic name (ICZN 2000). Creation of a Plectropomini identifies the outgroup to the grammistin lineages and restores monophyletic status to the rest of the Epinephelini. Additional taxa, characters from a locus slower in mutation rate, and larval character states for Variola are needed to resolve intertribe relationships within the Epinephelinae. The genus Variola contains only two described species, Variola albomarginata and V. louti, both easily recognized at the generic level by an oblong body form and lunate caudal tail, hence lyretails as their common name. The species lack larval descriptions, but differ in adult and juvenile coloration and relative pelvic-fin length, with V. albomarginata often mistaken for the more common V. louti (Heemstra and Randall 1993). Only one V. albomarginata was available in this study, and comparisons with 6 specimens of V. louti (2 were presumptive LDHA6 heterozygotes) had surprising results. The V. albomarginata specimen shares a cyt b haplotype with two specimens of V. louti. One of the V. louti shares a LDHA6 allele with one specimen of Plectropomus laevis, while differing from the V. albomarginata matching its haplotype by a single transition and 18 indels. This LDHA6 allele was present in several species of Epinephelus, causing one V. louti specimen to be removed from basal epinephelines in the intron gene tree
62 (Figure 3-4). Thus both gene trees detected deep lineage splits within this genus, but the splits were not consistent among specimens at the different loci. The Plectropomus mtDNA haplotypes were monophyletic, and shared a common ancestry with the Saloptia powelli haplotype observed in two specimens from the Marshall Islands (western central Pacific). LDHA6 alleles were not available from either specimen, but the strong support (93% probability) for this relationship at one locus concurs with other evidence. Saloptia and Plectropomus also possess a set of osteologic and meristic features that led Smith (1964) and Leis (1986) to suggest that these Indo-Pacific fishes are sister taxa. Both genera are restricted to the Indo-Pacific, although the depth range (140 to 367 m) could allow Saloptia to have a distribution broader than currently recorded. Another deepwater monotypic genus, Gonioplectrus, was not available for analysis, but larval morphology placed this genus intermediate between Plectropomus and Cephalopholis. Further analyses of DNA sequences and larval morphology are necessary to resolve the evolutionary history of these genera with soapfishes and groupers. Conclusions and Recommendations The cytochrome b locus, and to a lesser extent the LDHA6 intron, contained enough variation to evaluate relationships among epinepheline genera and thereby test the monophyly of tribe Epinephelini. A parallel exists between molecular and morphological data at many points in the gene trees. In addition to the agreements between alternate sources of data, both types of characters present problems for highly divergent taxa. Anyperodon leucogrammicus and Epinephelus itajara are notable examples of extreme morphological and ecological divergence within lineages considered here. A. leucogrammicus is extremely compressed and arguably the ultimate
63 morphological adaptation to exclusive piscivory. In contrast, most Epinephelus spp. consume some combination of fish, crabs, and small crustaceans. The gigantic size of the broad-headed E. itajara may restrict larger specimens to a diet comprised almost entirely of lobsters (Randall 1967). Both of these taxa are or have been in a separate genus of uncertain affinity, and this study has done little to resolve the situation. In the case of E. itajara, however, reconsideration of the subgenus Promicrops requires a study be made of E. lanceolatus, the Indo-Pacific member of the former genus. Two genera not examined at both loci deserve mention. Leis (1986) believed the larval morphology of Gonioplectrus to be intermediate between Plectropomus and Cephalopholis; making the placement of this genus within a tribe uncertain. The relative position of genus Paranthias may be easier to predict: these two species were noted by Craig et al. (2001) to have little genetic divergence at 16S from Cephalopholis fulvus. Given the uncommon but well-documented hybridization between these species (Bostrom et al. 2000), the inclusion of Paranthias within Cephalopholis may not be surprising. The strength of intrageneric relationships varies widely and species phylogenies are discussed in the Appendix. Obviously more data should be collected from both taxa and as many other species as possible to narrow the level of divergence and reconstruct their evolutionary history. The large number of species not included in this dataset could reveal polyphyletic and paraphyletic groups within the gene trees as currently described. However, some taxonomic changes may be warranted with the accumulation of a consistent pattern from independent sources (Table 3-1, Figure 3-5). The consistent failure of three loci to construct a monophyletic Epinephelini warrants removal of the genera Plectropomus,
64 Saloptia, and Variola from the tribe. Evidence has been presented in this and other studies for a lack of monophyly for the tribe Epinephelini and the genera Epinephelus, Cephalopholis, Dermatolepis, and Plectropomus. The ecological and morphological differences that set Dermatolepis, Alphestes, Anyperodon, Saloptia, and E. itajara apart from their closest ancestors are extensive; their taxonomic status as genera and subgenera may require an evaluation of not only other genetic loci but a detailed study of life history behavior, ecology, and morphology. Such tests of the evolutionary hypothesis proposed here (a tribe Plectropomini ancestral to the soapfish tribes and the Epinephelini; Figure 3-5) provides a framework for testing diversification mechanisms in the marine tropics. A well-described evolutionary history affects more than taxonomic practice (Collette and Vecchione 1995). Evaluations of global fishery resources and ecosystem-wide effects of fishing (e.g., Pauly et al. 1998, Sadovy 2001, Vecchione et al. 2000) are based in part upon agency reports of â€œgrouperâ€ catch data. Designation of a Plectropomini could remove exploited coralgroupers and lyretails (both easily distinguished from epinephelins) to a separate category, refining the accuracy and relevance of fishery statistics in the Indo-Pacific. The quantification of biodiversity patterns has increased in popularity and scope, resulting in fine-scale maps of habitat and faunal diversity (e.g., Bellwood and Hughes 2001, McAllister et al. 1994, Moritz and Faith 1998, Roberts et al. 2001). The taxonomic unit considered for tropical reef assessment is not only species but families and subfamilies as well, providing criteria for identifying areas of endemism (De Grave et al. 2001, Gill and Kemp 2002, Knowlton 2001, Roberts et al. 2002). Phylogenetic accuracy in taxonomic schema is the prerequisite for accurate evaluation of such â€˜biodiversity hotspotsâ€™ and an understanding
65 of the evolutionary ecology of coral reef fishes (Bellwood and Wainwright 2002, Choat and Bellwood 1991).
66 Table 3-1. Four alternative classifications of grouper and coralgrouper genera Smith (1971) Johnson (1983) Heemstra and Randall (1991)c this study SERRANIDAE SERRANIDAE SERRANIDAE SERRANIDAE SERRANINAE SERRANINAE SERRANINAE SERRANINAE Serranini ANTHIINAE ANTHIINAE ANTHIINAE Anthiini EPINEPHELINAE EPINEPHELINAE NIPHONINAE Niphonini NIPHONINAE Niphonini EPINEPHELINAE Epinephelini EPINEPHELINAE Epinephelini Epinephelini Cephalopholis Paranthias incertae sedis incertae sedis Aethaloperca Mycteropercaa Anyperodon Plectropomus Paranthias Epinephelus Cromileptes Saloptia Epinephelus (Cephalopholisb) Epinephelus Gonioplectrus Alphestes (Promicrops) (Alphestes) Cephalopholis Dermatolepis (Epinephelus) (Epinephelus) Aethaloperca Mycteroperca (Alphestes) (Cephalopholisb) Gracila Epinephelusd (Dermatolepis) (Dermatolepis) Mycteroperca (Promicrops) (Promicrops) Epinephelus Anyperodon incertae sedis Gonioplectrus (Promicrops) Cromileptes Aethaloperca Gracila Alphestes Anyperodon Mycteropercaa Dermatolepis tribe nov.e Cromileptes Paranthias Triso Plectropomus Gonioplectrus Plectropomus Paranthias Saloptia Gracila Variola Cromileptes Variola Plectropomus Variola Saloptia incertae sedis Variola Gonioplectrus Gracila Triso GRAMMISTINAE Diploprionini Diploprionini Diploprionini Diploprionini Liopropomini Liopropomini Liopropomini Liopropomini Grammistini Grammistini Grammistini Grammistini Note: Names are listed in phylogenetic order (most ancestral to most derived). Parenthetical names indicate subgenera. a, The genus Mycteroperca included Mycteroperca (=Trisotropis) dermopterus until removed to the monotypic genus Triso by Randall et al. (1989). Cephalopholis (Aethaloperca) rogaa. b, The genus Cephalopholis included C. albomarginatus until its removal to the monotypic genus Gracila by Smith-Vaniz et al. (1988). c, Although inset generic names indicate hypotheses of relationship, authors did not perform phylogenetic analyses. d, This Epinephelus clade is the same as the Epinephelus (Hyporthodus) clade of Craig et al. (2001) including E. (Promicrops) itajara. The subgeneric status of Promicrops may be in question (see text). e, The sister clade to Epinephelini and ancestral to the soapfish tribes would be named as a tribe after its oldest member genus, Plectropomus Oken 1817 (ICZN 2001).
Table 3-2. Nucleotide composition of DNA sequence alignments from percoid fishes sampled at cytochrome b (cyt b) and LDHA intron 6 Locus Taxa Base Composition Phylogenetic Signal Alignment m % A % C % G % T P Ti:Tv Variable Pars Inf Cyt b (n = 665) All sequences 87 24.41 29.98 15.54 30.07 1.000 1.61 67:18:221 55:11:221 Epinephelinae 65 24.34 30.36 15.41 29.89 1.000 2.19 50:11:220 45: 7:219 (Epinephelini) 57 24.27 30.64 15.72 29.72 1.000 2.61 42: 9:219 35: 6:216 LDHA6 No Indels (n =116) All sequences 61 23.66 23.84 23.80 28.70 1.000 1.16 2 5 13 Epinephelinae 57 23.61 23.82 23.87 28.69 1.000 1.23 2 1 12 (Epinephelini) 44 23.05 24.05 24.41 28.49 1.000 1.44 0 1 0 All Sites (n =186) All sequences 61 24.23 23.40 23.36 29.01 1.000 â€” 5 8 Epinephelinae 57 25.31 23.40 23.20 28.08 1.000 â€” 1 7 (Epinephelini) 44 24.78 24.82 23.44 26.95 1.000 â€” 2 1 Indels Only (n =70) All sequences 61 â€” â€” â€” â€” â€” â€” 3 6 Epinephelinae 57 â€” â€” â€” â€” â€” â€” 9 4 (Epinephelini) 44 â€” â€” â€” â€” â€” â€” 0 9 67
68 Table 3-3. Results of a posteriori topology tests among 50% majority-rule consensus trees generated from epinephelin cytochrome b Cytochrome b Tree SH Likelihood Tests Method Model lnL P Minimum Evolution p-distance 13816.760 59.054 0.075 p BC 13959.037 201.331 0.000* Maximum Parsimony equal weighting 14093.645 335.939 0.000* Weighted Parsimony 4:1 tv:ti 14066.530 308.825 0.000* 4:1 tv:ti, 3:5:1 codon 13816.760 59.054 0.075 Bayesian Likelihood chain heat = 0.2 13764.102 6.396 0.710 chain heat = 0.4 13757.706 (best) Note: BC = bootstrap consensus after 500 iterations. Bayes consensus = consensus of all Bayesian trees generated at different powers. difference between given topology and shortest tree. P-values with an asterisk (*) indicate a tree significantly different by Shimodaira-Hasegawa testing from the most likely tree at a significance of = 0.05.
69 Table 3-4. Results of a posteriori topology tests among 50% majority-rule consensus trees generated from epinephelin LDHA6 introns LDHA6 Intron Tree SH Likelihood Tests Method Model Gaps lnL P Minimum p-distance N 1277.604 5.011 0.789 Evolution p BC N 1354.903 82.310 0.078 mean NJ O 1409.880 137.288 0.001* mean NJ BC O 1515.243 242.650 0.000* Maximum equal weighting N 1276.975 4.382 0.902 Parsimony equal weighting E 1274.885 2.293 0.912 equal weighting O 1742.813 470.220 0.000* Weighted 2:1 tv:ti N 1364.899 92.307 0.702 Parsimony 2:1 tv:ti E 1272.593 (best) 2:1 tv:ti, 5:1 indel W 1284.157 11.565 0.702 Bayesian chain heat = 0.2 N 1316.618 44.025 0.354 Likelihood chain heat = 0.4 N 1475.678 44.959 0.351 Note: BC = bootstrap consensus after 500 iterations. NJ = Neighbor-joining. Bayes consensus = consensus of all Bayesian trees generated at different powers. Treatment of gaps in the alignment is encoded as: N: not included in analysis, E: treated as 5th base with equal weighting as transitions, W: treated as 5th base weighted 5 times transitions, O: only gaps were analyzed (coded as presence/absence characters). : Difference between given topology and shortest tree. P-values with an asterisk (*) indicate a tree significantly different by Shimodaira-Hasegawa testing from the most likely tree at a significance of = 0.05.
70 Table 3-5. Results of topology tests among unconstrained gene trees and gene trees with a priori topological constraints Locus KHTests Enforced Constraint Source Length Diff P Result Cyt b mtDNA monophyletic Epinephelini Johnson (1983) 13914.356 156.649 <0.001* H0 rejected; Epinephelini paraphyletic 5 subgenera within Epinephelus Smith (1971) 13977.082 219.376 <0.001* H0 rejected; Smithâ€™s Epinephelus is paraphyletic Plectropomus is ancestral to Mycteroperca + Epinephelus Leis (1986) 13834.189 76.483 0.001* H0 rejected; rejection due to paraphyletic Epinephelus 16SrRNA mtDNA gene tree Craig et al. (2001) 14150.717 393.011 <0.001* H0 rejected; data not explained by 16S rRNA tree LDHA6 intron tree this study 15577.693 1819.987 <0.001* H0 rejected; one or both gene trees do not equal the species tree none this study 13757.706 (best) best explains cyt b data LDHA6 intron 5 subgenera within Epinephelus Smith (1971) 958 176 <0.001* H0 rejected as above Plectropomus is ancestral to Mycteroperca + Epinephelus Leis (1986) 865 83 <0.001* H0 rejected as above 16SrRNA mtDNA gene tree Craig et al. (2001) 967 185 <0.001* H0 rejected as above cyt b mtDNA gene tree this study 997 215 <0.001* H0 rejected; one or both gene trees do not equal the species tree none this study 782 (best) best explains intron data Note: Cyt b trees were generated by Bayesian analysis, LDHA6 intron trees were maximum parsimony cladograms including gaps as a fifth base. Lengths and differences in lengths (Diff) were calculated using likelihood scores (cyt b) or parsimony scores (LDHA6). The Johnson (1983) hypothesis (i.e., Epinephelini is monophyletic) could not be tested at LDHA6 intron due to lack of sequences from outgroup taxa. Asterisked values in table indicate significant difference at P < 0.05.
71 Figure 3-1. Phylogenetic hypotheses of epinephelin evolution. Hypotheses are based upon (A) adult meristics and body shape (Smith 1971), (B) larval development (Leis 1986), and (C) a 16S gene tree (Craig et al. 2001). Triangles indicate unresolved intrageneric relationships.
72 Figure 3-2. The most likely gene tree for cytochrome b, showing derived species only. Numbers indicate node probability (shown above nodes when >50%) or numbers of specimens (shown right of taxon name when n > 1). Gray vertical bars indicate groups of intraspecific haplotypes. Abbreviations: Alp = Alphestes, Any = Anyperodon, Cep = Cephalopholis, Cro = Cromileptes, Der = Dermatolepis, Epi = Epinephelus, Myc = Mycteroperca.
73 Figure 3-3. The most likely gene tree for cytochrome b, showing ancestral species only. Abbreviations follow those in Figure 3-2 and: Aet = Aethaloperca, Apo = Aporops, Cep = Cephalopholis, Dic = Dicentrarchus, Epi = Epinephelus, Hyp = Hypoplectrus, Lat = Lateolabrax, Par = Paralabrax, Ple = Plectropomus, Ryp = Rypticus, Sal = Saloptia, Ser = Serranus, Ste = Stereolepis, Var = Variola.
74 Figure 3-4. The most parsimonious epinephelin gene tree at LDHA6. Numbers at nodes indicate per cent node retention after resampling by 500 bootstrap iterations. Gray vertical bars indicate intraspecific haplotypes. Abbreviations are explained in Figures 3-2 and 3-3.
75 Figure 3-5. Hypothetical phylogeny of the groupers and their allies. Tree is an Adams consensus of the cytochrome b gene tree (Figures 3-2 and 3-3), the LDHA6 gene tree (Figure 3-4) and the 16S rRNA gene tree of Craig et al. (2001). Numbered nodes indicate the taxonomic limits of: 1 = Family Serranidae, 2 = Subfamily Serraninae, 3 = Subfamily Epinephelinae, 4 = Tribe Epinephelini, 5 = Tribe Plectropomini, 6 = Tribe Grammistini. All other abbreviations follow those in Figures 3-2 and 3-3.
76 CHAPTER 4 CONCLUSIONS The groupers and coralgroupers exhibit a wide array of molecular evolutionary patterns. Gene flow in epinephelines has been described among genera (Bostrom et al. 2000) and species within genera (E. nigritus and E. merra, this study; Glamuzina et al. 2001, van Herwerden et al. 2002) while other species show a distinct genetic break among regional samples (E. adscensionis, E. fasciatus, R. saponaceous, Plectropomus; this study). The consistent failure of three loci to construct a monophyletic Epinephelini warrants creation of the tribe Plectropomini, including at least the genera Plectropomus, Saloptia, and Variola. Evidence has been presented in this and other studies for a lack of monophyly for the tribe Epinephelini and the genera Epinephelus, Cephalopholis, Dermatolepis, and Plectropomus. The epinepheline species E. adscensionis and the grammistine Rypticus saponaceous were each monophyletic, but with deep divergences between allopatric lineages. The molecular genealogies described thus far are typified by having a few ancient lineages with recent and extensive cladogenesis (i.e., Category I phylogeographic pattern; Avise 2000). Thus, the epinephelines may be an example of episodic, or â€˜punctuated,â€™ speciation, with speciation rates varying between clades (e.g., between E. itajara and other groupers) and time periods (Jackson and Cheetham 1999, Schluter 2000). Examples of unequal evolutionary rates within monophyletic clades are common in speciose marine taxa (Caldara et al. 1996, Johns and Avise 1998, McMillan and Palumbi 76
77 1997, Streelman et al. 2002). However, the same patterns could arise under a model of constant diversification but with high extinction rates, with extinction by chance or selection of incipient lineages (Avise 2000). Future investigations of grouper larval morphology, dispersal ability, adult life history characteristics, and habitat requirements could utilize these data to address the role of the Indo-Pacific triangle in reef fish evolution (Bellwood and Hughes 2001, Bernardi et al. 2000, Santini and Winterbottom 2002); the open population paradigm of coral reef fishes (Leis 2002, Mora and Sale 2002); the relationship among genetic stocks, fishery practice, and effective population size (Baker et al. 1995, Smith 1994, Smith et al. 1991); the evolution and retention of microsatellite and intron sequences (Dimmick 2001, Orti et al. 1997, Rico et al. 1996); and creation of species/region forensic database for monitoring trade in grouper products (Baker et al. 1996, Sweijd et al. 2000).
78 APPENDIX COMMENTS ON GROUPER SPECIES RELATIONSHIPS Species Relationships within a Reduced Epinephelini The de facto incomplete taxon sampling involved when delineating a large species complex prevents rigorous testing of species limits in this study. However, some relationships among haplotypes and genotypes are well supported and concordant with other lines of evidence. Perhaps more interesting to the evolutionary biologist are those cases in which no concordance among mitochondrial, nuclear, larval or adult morphological data was present. Detailed accounts of major groupings and problematic taxa are given below in order of divergence from outgroup taxa. The Hind Clade Five species of Cephalopholis were examined, and haplotypes from each formed monophyletic groups in two distinct and divergent lineages. A well-resolved cyt b clade comprised of the East Atlantic C. taeniops as ancestral sister taxon of the West Atlantic C. fulvus and the Indo-West Pacific C. sexmaculata. Cephalopholis fulvus and C. taeniops were sister taxa in the LDHA6 gene tree as well. A second lineage within Cephalopholis observed at both cyt b and LDHA6 contains C. nigri and C. cruentata as sister taxa. At 16S, C. cruentata was moderately supported (76% bootstrap support) as ancestral to a clade including C. fulvus. A sample of A. rogaa from Hong Kong was placed into the Cephalopholis clade, but at the LDHA6 locus, this specimen was placed into an unresolved polytomy with Plectropomus, thus the evolutionary relationships of this genus remains uncertain. 78
79 The True Grouper Clade The data presented in this study identify a sharp genetic divergence among the coralgroupers, lyretails, and hinds from other sampled epinephelin genera, confirming general hypotheses based upon larval morphology and 16S gene trees. Leis (1986) noted an increase in dorsal-fin spine number from 8 directly forming spines (Plectropomus), to 8+1=9 directly+indirectly forming spines (Cephalopholis) to the 9 to 11 spines of Epinephelus and Mycteroperca. The node is well supported at cyt b (P=98%), LDHA6 (P~100%), and 16S (87% bootstrap support; Craig et al. 2001). The more ancestral genera of Cephalopholis, Saloptia, and Variola can be generalized as oblong piscivores with a distribution nearly restricted to Indo-Pacific waters. The remaining genera sampled in this study have a relatively deeper body form and considerable representation in the Atlantic. Weaver (1996) suggested that serranid body form may be the result of a coupling between prey size and energy yield with gape limitation and predator avoidance. Smaller size (like the 30 to 40 cm Cephalopholis) increases reliance upon sprint speed, visual acuity, and dependence upon reefs for shelter and predator avoidance. An increased body size means a lessening of gape limitation, perhaps resulting in capture of prey items with higher energy yield. While it is difficult to distinguish if the inclusion of more demersal, hard-shelled prey was a result or a progenitor of an adaptive radiation, the mtDNA data lend some support to the existence of such a morphological spectrum in epinephelins. Of course, exceptions can be named: the derived Anyperodonâ€™s elongate body form would have to be encoded as a derived compression while the crustacean-eating C. sonnerati, whose body depth is nearly equal to its standard length, would represent convergence upon an Epinephelus-like form from within a more primitive lineage.
80 The Haifa and Leather Bass Clades The Haifa grouper E. haifensis unambiguously roots (P = 100%) a monophyletic group of five groupers in the cyt b gene tree. Epinephelus flavolimbatus is a sister taxon to E. niveatus, E. niphobles, and a haplotype shared between the west Atlantic E. nigritus and the Indo-Pacific E. merra. The relationship among E. nigritus, E. niphobles and E. niveatus is mirrored at 16S (Craig et al. 2001). Members of the Haifa clade differed from each other by 1 to 7 substitutions at cyt b, but all shared seven deletions in LDHA6 relative to all other sequences, creating a strongly divergent clade intermediate between Cephalopholis intron alleles and true grouper intron alleles. However, two other E. merra and one other E. nigritus did not possess these deletions and differed by only 1 to 3 substitutions. Smith (1971) synonymized E. niphobles and E. niveatus because they have identical juvenile coloration and adult meristics. Heemstra and Randall (1993) disputed the taxonomic decision, citing slight differences in eye and pectoral fin morphology, but did not deny a potentially close relationship. The close relationship between the E. niphobles/E. niveatus pair and the west Atlantic E. nigritus is more unexpected. Epinephelus nigritus is the only member of the Haifa clade to have X instead of XI dorsal-fin spines, and to have 4 to 5 rows of teeth instead of 2 to 3 rows on the midlateral of the lower jaw. Leis (1986) commented upon the confusing nature of dorsal-fin spine counts, noting that direct formation of 8 or 9 spines are reliable characters for unifying epinephelin genera, but that spines beyond the count of 9 may form directly (orthologous) or form by fusion of fin rays (paralogous). The number of tooth rows is notable because juvenile E. nigritus possess 2 to 3 midlateral tooth rows and develop additional rows upon settlement.
81 The closest sister taxon to the Haifa clade in this study is a strongly supported (at cyt b and 16S) clade including the leather bass Dermatolepis dermatolepis and two of the three described species of Alphestes. Both genera possess morphological novelties that have made past attempts at phylogenetic placement problematic (Heemstra and Randall 1993). The genetic divergence among Alphestes, Dermatolepis, and Epinephelus is large, thus the leather bass clade may represent early specializations outside most species in the true grouper complex. The Mycteroperca Clade The species of Mycteroperca form a well-supported monophyletic group at cyt b and at least partially at LDHA6. The genus is restricted to the Atlantic and East Pacific, and surely represents a radiation from a single lineage predating the closure of the Panamanian Seaway 3.1 to 3.7 MYA (Coates and Obando 1996). Only one East Pacific species, M. rosacea, was available for analysis. The cyt b M. rosacea haplotype differed from M. bonaci by a single transition; these differed from two M. venenosa haplotypes by 28 to 29 substitutions. The placement of M. bonaci and M. venenosa as sister taxa is supported at 16S, whereas M. tigris intron alleles are sister taxa to M. venenosa intron alleles. Unlike 16S, the cyt b Mycteroperca clade did not include Anyperodon (see discussion below). The most divergent Mycteroperca sequence of those sampled in this study was M. fusca. Specimens of M. fusca from the Gulf of Mexico and the Atlantic coast of Florida shared a haplotype with a specimen from the Azores, the first record of M. fusca in these Mid-Atlantic islands (Morato et al., in review). The cyt b haplotype and two sister LDHA6 alleles were most closely related to specimens of Epinephelus marginatus from the Azores in the central Atlantic. The 56 ti and 16 tv difference at cyt b resulted in a single nonsynonymous amino acid substitution, and the LDHA6 alleles
82 differed by 1 to 2 ti, 2-4 tv, and 0 indels. With such sequence similarity, E. marginatus and M. fusca specimens were sister taxa in NJ, MP, WP, and Bayesian likelihood trees at both loci despite meristic and morphological differences between the two genera (Heemstra and Randall 1993). The E. marginatus and M. fusca lineage may also include the Mediterranean species E. caninus, based upon analyses of 431 bp of cyt b (T. Maggio, Univ. de Palermo, pers. comm.). The American Clade Although Smithâ€™s (1971) assumption of monophyly among all American groupers is not supported (this study, Craig et al. 2001), six Epinephelus species from the West Atlantic and East Pacific form a well-supported (P=94%) monophyletic lineage in the cyt b gene tree. Epinephelus striatus, E. morio, and E. guttatus show identical relationships at cyt b, LDHA6, and 16S. Johnson and Keener (1984) were unable to distinguish the larvae of E. morio and E. guttatus from each other or E. drummondhayi. These three species have identical historical distributions except that E. guttatus is absent from southern Brazil, and form a sister clade to a clade comprised of Epinephelus adscensionis, E. labriformis, and E. analogus. Although E. adscensionis and E. analogus have been proposed as geminate species across the Panamanian Isthmus (Heemstra and Randall, 1993), it is the East Pacific E. labriformis that separates the West Atlantic E. adscensionis from the East Pacific by 9.5% sequence divergence. Given the total overlap among geographic and depth distributions (Heemstra and Randall 1993), nearly identical diets (Randall 1967), and microsympatry at the same fishing grounds (Heemstra and Randall 1993), the American clade therefore presents an opportunity for studying speciation tempo and mode.
83 The Pacific Grouper and Promicrops Group The West Pacific species E. awoara, one of the relatively few groupers in the East China Sea and a species of increasing importance in grouper aquaculture (Sadovy 2000, Sadovy and Cornish 2000), could not be unambiguously assigned a position among true grouper clades. A set of cyt b sequences was aligned with moderate support at cyt b (P = 79%) that includes a strongly supported set of four species and two highly divergent taxa. Epinephelus caerulopunctatus and E. tauvina positioned as sister taxa to E. polyphekadion and C. altivelis. The pairs are joined with strong support in the cyt b DNA alignment (P = 96%) but are not joined upon conversion to an amino acid NJ phylogram. The monotypic genus Cromileptes has a highly specialized cranial morphology with a dorsal-fin shape, position, and spine count unique in Indo-Pacific groupers. The pairing of E. polyphekadion as a sister taxon of Cromileptes may be the result of long-branch attraction, although the relationship is supported after alignment translation. The same might be said for the strong support for joining the elongate exclusive piscivore E. tauvina and the deeper-bodied and crustacean-feeding E. caerulopunctatus. The monotypic genus Anyperodon was placed as a sister taxon of E. itajara at cyt b, a sister of Cromileptes at LDHA6, and within the Mycteroperca clade at 16S. The three loci, together with an advanced pterygiophore arrangement (Smith-Vaniz et al. 1988), are concordant with a generic placement as a derived true grouper, but otherwise its phylogenetic position cannot be determined with reliability. The genus Promicrops included two species with a unique arrangement of lateral line tubules and gigantic size (up to at least 2.5 m and 320 kg). Smith (1971) reduced Promicrops to a subgenus of Epinephelus, a decision supported by later workers, although Heemstra and Randall (1993) noted the extreme specialization present in the
84 genus. Johnson and Keener (1984) noted that E. itajara larvae are characterized by having pigmentation recorded from Mycteroperca and a Cephalopholis species. Two samples of Atlantic E. itajara shared identical LDHA6 alleles with three unique deletions and a mean sequence divergence of 3.7% from various true grouper lineages (including Anyperodon and E. polyphekadion). Epinephelus itajara intron allele was the ancestral sequence to most Epinephelus and Mycteroperca sequences. The divergence between E. itajara at cyt b, however, was one of the greatest seen within the true groupers (mean d=15.7% between E. itajara and E. tauvina). The cyt b divergence was overwhelmingly comprised of nonsynonymous substitutions, and translated to 15 to 21 amino acid substitutions between either E. itajara haplotype and any other amino acid sequence in the Epinephelinae. Obviously additional loci and morphological characters are needed before the systematic relationships of the subgenus Promicrops, or the systematic status of Epinephelus (Promicrops) itajara can be determined. Final Thoughts Taxonomies may reflect ease of diagnosis, an attempt to reconstruct phylogenetic history, or a need for stability. Usually in practice, taxonomies are some combination of all three (Stevens 1992). The de facto incomplete taxon sampling involved when delineating a large species complex prevents rigorous testing of species limits in this study. At first glance many binomials could be questioned on the basis of the mtDNA gene trees alone. Gilles et al. (2000) suggested that E. marginatus is comprised of cryptic species, and Carlin et al. (in press) identified a similar genetic divergence within E. adscensionis. Yet it is unknown whether the alternate phylogeographic units with symparapatric distribution fit definitions as different species. Alternate species concepts may change the taxonomic label of a unit, but this does not change its cladistic
85 relationship. Even as phylogenetic species, confirmation at multiple loci is preferable to reliance upon a single locus (Slowinski and Page 1999). Are these organisms failing to interbreed entirely? Do they form cohesive lineages at multiple loci across their range and across generations? Dupr (1992) has noted that evolutionary biologists historically state belief in a species concept without providing the means to test its assumptions. The groupers and coralgroupers possess a wide range of features which, examined in the context of an accurate genealogical framework, can provide the means of testing speciation processes.
86 LIST OF REFERENCES Allegrucci, G., A. Caccone, and V. Sbondoni. 1999. Cytochrome b sequence divergence in the European sea bass (Dicentrarchus labrax) and phylogenetic relationships among some Perciformes species. J. Zool. Syst. Evol. Res. 37:149-156. Allen, G. R., and D. R. Robertson. 1997. An annotated checklist of the fishes of Clipperton atoll, tropical eastern Pacific. Rev. Biol. Trop. 45:813-843. Amos, B., and A. R. Hoelzel. 1991. Long-term preservation of whale skin for DNA analysis. Rep. Int. Whal. Comm. Spec. Iss. 13:99-103. Anderson, S., A. T. Bankier, B. G. Barrell, M. H. L. de Bruijin, A. R. Coulson , J. Drouin, I. C. Eperon , D. P. Nierlich, B. A. Roe, F. Sanger, P. H. Schreier, A. J. H. Smith, R. Staden, and I. G. Young. 1981. Sequence and organization of the human mitochondrial genome. Nature 290:457-465. Avise, J. C. 2000. Phylogeography: The history and formation of species. Harvard University Press, Cambridge, Massachusetts. Baker, C. S., F. Cipriano, and S. R. Palumbi. 1996. Molecular genetic identification of whale and dolphin products from commercial markets in Korea and Japan. Mol. Ecol. 5:671-685. Baker, C. S., A. Perry, G. K. Chambers, and P. J. Smith. 1995. Population variation in the mitochondrial cytochrome b gene of the orange roughy Hoplostethus atlanticus and the hoki Macruronus novaezelandiae. Mar. Biol. 122:503-509. Baldwin, C. C., and G. D. Johnson. 1993. Phylogeny of the Epinephelinae (Teleostei: Serranidae). Bull. Mar. Sci. 52:240-283. Ball, A. O., G.R. Sedberry, M. Zatcoff, R.W. Chapman, and J.L. Carlin. 2000. Population structure of wreckfish (Polyprion americanus) determined using microsatellite genetic markers. Mar. Biol. 137:1077-1090. Banford, H. M., E. Bermingham, B. B. Collette, and S. S. McCafferty. 1999. Phylogenetic systematics of the Scomberomorus regalis (Teleostei: Scombridae) species complex: molecules, morphology, and biogeography of Spanish mackerels. Copeia 1999:596-613. 86
87 Bellwood, D. R., and T. P. Hughes. 2001. Regional-scale assembly rules and biodiversity of coral reefs. Science 292:1532-1534. Bellwood, D. R., and P. C. Wainwright. 2002. The history and biogeography of fishes on coral reefs. Pp. 5-32. in: Coral reef fishes: dynamics and diversity in a complex ecosystem. P. F. Sale (ed.). Academic Press Inc, San Diego, California. Bernardi, G., D. R. Robertson, K. E. Clifton, and E. Azzurro. 2000. Molecular systematics, zoogeography, and evolutionary ecology of the Atlantic parrotfish Sparisoma. Mol. Phy. Evol. 15:292-300. Birt, T. P., V. L. Friesen, R. D. Birt, J. M. Green, and W. S. Davidson. 1995. Mitochondrial DNA variation in Atlantic capelin, Mallotus villosus: a comparison of restriction and sequence analyses. Mol. Ecol. 4:771-776. Bhlke, J. E., and C. C. G. Chaplin. 1993. Fishes of the Bahamas and adjacent tropical waters, second edition. University of Texas Press, Austin, Texas. Bonilla, J., W. Senior, J. Bugden, O. Zafiriou, and R. Jones. 1993. Seasonal distribution of nutrients and primary productivity on the eastern continental shelf of Venezuela as influenced by the Orinoco River. J. Geophys Res. 98:2245-2257. Bostrom, M. A., B. B. Collette, B. E. Luckhurst, K. S. Reece, and J. E. Graves. 2000. Hybridization between two serranids: the coney Cephalopholis fulva and the creole-fish Paranthias furcifer, at Bermuda. Fish. Bull. 100:651-661. Bowen, B. W., A. L. Bass, L. A. Rocha, W. S. Grant, and D. Ross Robertson. 2001. Phylogeography of the trumpetfishes (Aulostomus): ring species complex on a global scale. Evolution 55:1029-1039. Bowen, B. W., and W. S. Grant. 1997. Phylogeography of the sardines (Sardinops spp.): Assessing biogeographic models and population histories in temperate upwelling zones. Evolution 51:1601-1610. Braun, E. L., and R. T. Kimball. 2002. Examining basal avian divergences with mitochondrial sequences: model complexity, taxon sampling, and sequence length. Syst. Biol. 51:614-624. Briggs, J. C. 1974. Marine biogeography. McMillan and Sons, New York, New York. Briggs, J. C. 1995. Global biogeography. Elsevier, Amsterdam, Netherlands. Brooks, D. R., and D. A. McLennan. 2002. The nature of diversity: an evolutionary voyage of discovery. University of Chicago Press, Chicago, Illinois.
88 Brothers, E. B., and R. E. Thresher. 1985. Pelagic duration, dispersal, and the distribution of Indo-Pacific coral reef fishes. Pages 53-69 in M. L. Reaka, eds. The Ecology of Deep and Shallow Coral Reefs. NOAA Special Publications, Rockville, Maryland. Bucklin, A., and P. H. Wiebe. 1998. Low mitochondrial diversity and small effective population sizes of copepods Calanus finmarchius and Nannocalanus minor: Possible impact of climatic variation during recent glaciation. J. Hered. 89:383-392. Caldara, F., L. Bargelloni, L. Ostellari, E. Penzo, L. Colombo, and T. Patarnello. 1996. Molecular phylogeny of grey mullets based on mitochondrial DNA sequence analysis: Evidence of a differential rate of evolution at the intrafamily level. Mol. Phylo. Evol. 6:416-424. Canatore, R., M. Roberti, G. Pesole, A. Ludovico, F. Milella, M. N. Gadaleta, and C. Saccone. 1994. Evolutionary analysis of cytochrome b sequences in some Perciformes: evidence for a slower rate of evolution than in mammals. J. Mol. Evol. 39:589-597 Carlin, J. L. 2002. Molecular evolution in the cytochrome b gene of epinepheline fishes (Percoidei: Serranidae). Proc. World Cong. Biol. Fish, Vancouver, British Columbia, Canada. Carlin, J. L., D. R. Robertson, and B. W. Bowen. In Press. Ancient divergences in the serranids Epinephelus adscensionis and Rypticus saponaceous. Marine Biology. Cavalcanti, M. J., L. R. Monteiro, and P. R. D. Lopes. 1999. Landmark-based morphometric analysis in selected species of serranid fishes (Perciformes: Teleostei). Zool. Stud. 38:287-294. Coates, A. G., and J. A. Obando. 1996. The geologic evolution of the central American isthmus. Pp. 21-56 in: Evolution and environment in tropical America. J. B. C. Jackson, A. F. Budd, A. G. Coates (eds.). University of Chicago Press, Chicago, Illinois. Choat, J. H., and D. R. Bellwood. 1991. Reef fishes: their history and evolution, p. 39-66 in: The ecology of fishes on coral reefs. P. F. Sale (ed.). Academic Press Inc., San Diego, California. Colborn, J., R. E. Crabtree, J. B. Shaklee, E. Pfeiler, and B. W. Bowen. 2001. The evolutionary enigma of bonefishes (Albula spp.): cryptic species and ancient separations in a globally distributed shorefish. Evolution 55:807-820.
89 Coleman, F. C., C. C. Koenig, G. R. Huntsman, J. A. Musick, A. M. Eklund, J. C. McGovern, R. W. Chapman, G. R. Sedberry, and C. B. Grimes. 2001. Long-lived reef fishes: the grouper-snapper complex. Fisheries 25:14-21. Collette, B. B., and K. Rtzler. 1977. Reef fishes over sponge bottoms off the mouth of the Amazon River. Proc. Third Intl. Coral Reef Symp.:305-309. Collette, B. B., and M. Vecchione. 1995. Interactions between fisheries and systematics. Fisheries 20:20-25. Collins, L. S., A. G. Coates, W. A. Berggren, M. Aubry, J. Zhang. 1996. The late Miocene Panama isthmian strait. Geology 24:687-690. Courtenay Jr., W. R. 1967. Atlantic fishes of the genus Rypticus (Grammistidae). Proc. Acad. Nat. Sci. Phil. 119:241-293. Courtenay Jr., W. R. 1970. The R/V Pilsbury deep-sea biological expedition to the Gulf of Guinea, 1964-65: Soapfishes of the genus Rypticus from Fernando Poo. Stud. Trop. Ocean. 4:276-280. Cracraft, J. 1983. Species concepts and speciation analysis. Cur. Ornith. 1:159-187. Cracraft, J. 1987. Species concepts and the ontology of evolution. Biol. Phil. 2:329-346. Craig, M. T., D. J. Pondella II, J. P. C. Franck, and J. C. Hafner. 2001. On the status of the serranid fish genus Epinephelus: Evidence for paraphyly based on 16S rDNA sequence. Mol. Phyl. Evol. 19:121-130. Cunningham, C. W., H. Zhu, and D. M. Hillis. 1998. Best-fit maximum likelihood models for phylogenetic inference: empirical tests with known phylogenies. Evolution 52:978-987. De Grave, S. 2001. Biogeography of Indo-Pacific Pontoniinae (Crustacea, Decapoda): a PAE analysis. J. Biogeog. 28:1239-1253. Dimmick, W. W. 2001. Spliceosomal introns and fish phylogeny: a critical reanalysis. Copeia 2001:536-541. Dolphin, K., R. Belshaw, C. D. Orme, and D. L. J. Quike. 2000. Noise and incongruence: Interpreting results of the incongruence length difference test. Mol. Phylog. Evol. 17:401. Domeier, M. L., and P. L. Colin. 1997. Tropical reef fish spawning aggregations: defined and reviewed. Bull. Mar. Sci. 60:698-726.
90 Dudgeon, C. L., N. Gust, and D. Blair. 2000. No apparent genetic basis to demographic differences in scarid fishes across continental shelf of the Great Barrier Reef. Mar. Biol. 137:1059-1066. Dupr, J. 1992. Species: theoretical contexts. Pp. 312-317 in: Keywords in evolutionary biology. E. F. Keller and E. A. Lloyd (eds.). Harvard University Press, Cambridge, Massachusetts. Excoffier, L., P. E. Smouse, and J. M. Quattro. 1992. Analysis of molecular variance inferred from metric distances among DNA haplotypes: application to human mitochondrial DNA restriction data. Genetics 131:479-491. Farris, J. S., M. Kallersj, A. G. Kluge, and C. Bult. 1994. Testing significance of incongruence. Cladistics 10:315. Farris, J. S., M. Kallersj, A. G. Kluge, and C. Bult. 1995. Constructing a significance test for incongruence. Syst. Biol. 44:570. Fral, J. P. 2000. How useful are the genetic markers in attempts to understand and manage marine biodiversity? J. Exp. Mar. Biol. Ecol. 268:121-145. Fischer, W., G. Bianchi, and W. B. Scott, eds. 1981. FAO species identification sheets for fishery purposes; Eastern Central Atlantic; fishing areas 34, 47 (in part). Food and Agricultural Organization of the United Nations, Rome, Italy. Fleischer, R. C., C. E. McIntosh, and C. L. Tarr. 1998. Evolution on a volcanic conveyor belt: using phylogeographic reconstructions and K-Ar-based ages of the Hawaiian Islands to estimate molecular evolutionary rates. Mol. Ecol. 7:533-545. Floeter, S. R., R. Z. P. Guimaraes, L. A. Rocha, C. E. L. Ferreira, C. A. Rangel, and J. L. Gasparini. 2000. Geographic variation in reef-fish assemblages along the Brazilian coast. Glob. Ecol. Biog. 10:423-432. Floeter, S. R., and J. L. Gasparini. 2000. The southwestern Atlantic reef fish fauna: composition and zoogeographic patterns. J. Fish Biol. 56:1099-1114. Floeter, S. R., J. L. Gasparini, L. A. Rocha, R. Z. P. Guimares, C. E. L. Ferreira, C. A. Rangel, B. M. Feitoza, and G. W. Nunan. 2001. Brazilian reef fish fauna: checklist and remarks. BioBase Project: http://www.biobase.org/BCF/Index.html. Gill, A. C., and J. M. Kemp. 2002. Widespread Indo-Pacific shore-fish species: a challenge for taxonomists, biogeographers, ecologists, and fishery and conservation managers. Env. Biol. Fishes 65:165-174.
91 Gilles, A., A. Miquelis, J.-P. Quignard, and E. Faure. 2000. Molecular phylogeography of western Mediterranean dusky grouper Epinephelus marginatus. C. R. Acad. Sci./Life Sciences 323:195-205. Glamuzina, B., N. Glavic, B. Skaramuca, V. Kozul, P. Tutman. 2001. Early development of the hybrid Epinephelus costae female x E. marginatus male. Aquaculture 198:55-61. Gosline, W. A. 1960. A new Hawaiian percoid fish, Suttonia lineata, with a discussion of its relationships and a definition of the family Grammistidae. Pac. Sci. 14:28-38. Gosline, W. A. 1966. The limits of the fish family Serranidae, with notes on other lower percoids. Proc. Calif. Acad. Sci. 33:91-112. Grant, W. S., and B. W. Bowen. 1998. Shallow population histories in deep evolutionary lineages of marine fishes: Insights from sardines and anchovies and lessons for conservation. J. Hered. 89:415-426. Graves, J. E. 1995. Conservation genetics of fishes in the pelagic marine realm. Pp. 335-366 in: Conservation genetics: case histories from nature. J. C. Avise and J. L. Hamrick (eds). Chapman and Hall, New York, New York. Graves, J. E. 1998. Molecular insights into the population structures of cosmopolitan marine fishes. J. Hered. 89:427-437. Grove, J. S., and R. J. Lavenberg. 1997. The fishes of the Galapagos Islands. Stanford University Press, Stanford, CA. Guimares, R. Z. P. 1999. Revision, phylogeny and comments on biogeography of soapfishes of the genus Rypticus (Teleostei : Serranidae). Bull. Mar. Sci. 65:337-379. Haig, S. M. 1998. Molecular contributions to conservation. Ecology 79:413-425. Harpending, R. C. 1994. Signature of ancient population growth in a low-resolution mitochondrial DNA mismatch distribution. Hum. Biol. 66:591-600. Haug, G. H., and R. Tiedemann. 1998. Effect of the formation of the Isthmus of Panama on Atlantic Ocean thermohaline circulation. Nature 393:673-676. Hedgecock, D. 1994. Does variance in reproductive success limit effective population sizes of marine organisms? Pp. 122-134 in: Genetics and evolution of marine organisms. A. R. Beaumont (ed.). Chapman and Hall, New York, NY.
92 Heemstra, P. C., and J. E. Randall. 1993. Groupers of the world (Family Serranidae, subfamily Epinephelinae) : an annotated and illustrated catalogue of the grouper, rockcod, hind, coral grouper and lyretail species known to date. Food and Agriculture Organization of the United Nations, Rome. Hewitt, G. M. 2001. Speciation, hybrid zones and phylogeography or seeing genes in space and time. Mol. Ecol. 10:237-549. Hillis, D. M., C. Moritz and B. K. Mable. 1996. Molecular systematics. Sinauer Associates, Sunderland, Massachusetts. Huelsenbeck, J. P., and F. Ronquist. 2001. MrBayes version 2.01. Bayesian Inference of Phylogeny. Univ. of Rochester and Uppsala Univ., Rochester, New York. Humann, P. 2002. Reef fish identification: Florida Caribbean Bahamas. New World Publications Inc., Jacksonville, Florida. International Commission on Zoological Nomencalture (ICZN). 2000. The international code of zoological nomenclature. International Trust for Zoological Nomenclature, London, United Kingdom. International Marinelife Alliance (IMA) Indonesia. 2001. Reef fisheries exploitation and trade in Indonesia. International Marinelife Alliance, Bogor, Indonesia. Irwin, D. M., T. D. Kocher, and A. C. Wilson. 1991. Evolution of the cytochrome b gene of mammals. J. Mol. Evol. 32:128-144. Jackson, J. B. C., and A. H. Cheetham. 1999. Tempo and mode of speciation in the sea. TREE 14:72-77. Johannes, R. E. 1978. Reproductive strategies of coastal marine fishes in the tropics. Env. Biol. Fish 3:65-84. Johns, G. C., and J. C. Avise. 1998. A comparative summary of genetic distances in the vertebrates from the mitochondrial cytochrome b gene. Mol. Biol. Evol. 15:1481-1490. Johnson, G. D. 1983. Niphon spinosus: A primitive epinepheline serranid, with comments on the monophyly and intrarelationships of the Serranidae. Copeia 1983:777-787. Johnson, G. D. 1988. Niphon spinosus: A primitive epinepheline serranid: corroborative evidence from the larvae. Jap. J. Ichthyol. 35:7-18. Johnson, G. D., and P. Keener. 1984. Aid to identification of American grouper larvae. Bull. Mar. Sci. 34:106-134.
93 Keigwin, L. D. 1982. Isotopic paleoceanography of the Caribbean and East Pacific: role of Panama uplift in late Neogene time. Science 217:350-352. Kendall, A. W. Jr. 1976. Predorsal and associated bones in serranid and grammistid fishes. Bull. Mar. Sci. 26:585-592. Kishino, H., and M. Hasegawa. 1989. Evaluation of the maximum likelihood estimate of the evolutionary tree topologies from DNA sequence data, and the branching order in Hominoidea. J. Mol. Evol. 29:170-179. Knowlton, N. 2000. Molecular genetic analyses of species boundaries in the sea. Hydrobiologia 420:73-90. Knowlton, N. 2001. The future of coral reefs. Proc. Nat. Acad. Sci. USA 98:5419-5425. Lacson, J. M. 1992. Minimal genetic variation among samples of six species of coral reef fishes collected at La Parguera, Puerto Rico, and Discovery Bay, Jamaica. Mar. Biol. 112:327-331. Lacson, J. M., V. M. Riccardi, S. W. Calhoun, and D. C. Morizot. 1989. Genetic differentiation of bicolor damselfish (Eupomacentrus partitus) populations on the Florida Keys. Mar Biol 103:445-451. Lavery, S., C. Moritz, and D. R. Fielder. 1996. Genetic patterns suggest exponential population growth in a declining species. Mol. Biol. Evol. 13:1106-1113. Leis, J. M. 1986. Larval development in four species of Indo-Pacific coral trout Plectropomus (Pisces: Serranidae: Epinephelinae) with an analysis of the relationships of the genus. Bull. Mar. Sci. 38:525-552. Leis, J. M. 1991. The pelagic stage of reef fishes: the larval biology of coral reef fishes. Pp. 183-230 in: The Ecology of Fishes on Coral Reefs. P. F. Sale (eds.). Academic Press Inc., San Diego, California. Leis, J. M. 2002. Pacific coral-reef fishes: the implications of behavior and ecology of larvae for biodiversity and conservation, and a reassessment of the open population paradigm. Env. Biol. Fishes 65:199-208. Lessios, H. A., B. D. Kessing, and J. S. Pearse. 2001. Population structure and speciation in tropical seas: global phylogeography of the sea urchin Diadema. Evolution 55:955-975. Lindeman, K. C., T. N. Lee, W. D. Wilson, R. Claro, and J. S. Ault. 2000. Transport of larvae originating in southwest Cuba and the Dry Tortugas: Evidence for partial retention in grunts and snappers. Proc. Gulf Carib. Fish. Inst. 52:253-278.
94 Longhurst, A. 1998. Ecological geography of the sea. Academic Press Inc., San Diego, California. Maier-Reimer, E., U. Mikolajewicz, and T. J. Crowley. 1990. Ocean general circulation model sensitivity experiment with an open American isthmus. Paleoceanography 5:349-366. McAllister, D. E., F. W. Schueler, C. M. Roberts, and J. P. Hawkins. 1994. Mapping and GIS analysis of the global distribution of coral reef fishes on an equal-area grid. Pp. 155-176 in: Mapping the diversity of nature. R. I. Miller (ed.). Chapman and Hall, London, United Kingdom. McMillan, W. O., and S. R. Palumbi. 1997. Rapid rate of control-region evolution in Pacific butterflyfishes (Chaetodontidae). J Mol Evol 45:473-484. Mindell, D. P., and C. E. Thacker. 1996. Rates of molecular evolution: phylogenetic issues and applications. Ann Rev Ecol Syst 27:279-303. Mora, C., and P. F. Sale. 2002. Are populations of coral reef fish open or closed? TREE 17:422-428. Morato, T., P. Afonso, and J. L. Carlin. 2003. First record of scamp, Mycteroperca phenax, in the north-eastern Atlantic. J. Fish Biol.: in review. Moritz, C., and D. P. Faith. 1998. Comparative phylogeography and the identification of genetically divergent areas for conservation. Mol. Ecol. 7:419-429. Musick, J. A., M. M. Harbin, S. A. Berkeley, G. H. Burgess, A. M. Ecklund, L. Findley, R. G. Gilmore, J. T. Golden, D. S. Ha, G. R. Huntsman, J. C. McGovern, S. J. Parker, S. G. Poss, E. Sala, T. W. Schmidt, G. R. Sedberry, H. Weeks, and S. G. Wright. 2000. Marine, estuarine, and diadromous fish stocks at risk of extinction in North America (exclusive of Pacific salmonids). Fisheries 25:6-30. Muss, A., D. R. Robertson, C. A. Stepien, P. Wirtz, and B. W. Bowen. 2001. Phylogeography of Ophioblennius: the role of ocean currents and geography in reef fish evolution. Evolution 55:561-572. National Marine Fisheries Service (NMFS). 2002. Fisheries of the United States, 2001. U. S. Department of Commerce-NOAA-NMFS, Silver Spring, Maryland. Near, T. J., J. C. Porterfield, and L. M. Page. 2000. Evolution of cytochrome b and the molecular systematics of Ammocrypta (Percidae : Etheostomatinae). Copeia 2000(3):701-711.
95 Needleman, S. G., and C. D. Wunsch. 1970. A general method applicable to the search for similarities in the amino acid sequence of two proteins. J. Mol. Biol. 48:443-453. Nei, M. 1987. Molecular evolutionary genetics. Columbia University Press, New York, NY. Neigel, J. E. 1997. A comparison of alternative strategies for estimating gene flow from genetic markers. Annu. Rev. Ecol. Syst. 28:105-128. Nichols, R. 2001. Gene trees and species trees are not the same. TREE 16:358-363. Norris, R. D. 2000. Pelagic species diversity, biogeography, and evolution. Paleobiology 26, Suppl. S:S236-S258. Orti, G., D. E. Pearse, and J. C. Avise. 1997. Phylogenetic assessment of length variation at a microsatellite locus. Proc. Nat. Acad. Sci. USA 94:10745-10749. Palumbi, S. R. 1996. What can molecular genetics contribute to marine biogeography? An urchin's tale. J. Exp. Mar. Biol. Ecol. 203:75-92. Pamilo, P., and M. Nei. 1988. Relationships between gene trees and species trees. Mol. Biol. Evol. 5:568-583. Pauly, D., V. Christensen, J. Dalsgaard, R. Froese, and F. Torres, Jr. 1998. Fishing down marine food webs. Science 279:860-863. Planes, S., F. Bonhomme, and R. Galzin 1993. Genetic structure of Dascyllus aruanus populations in French Polynesia. Mar. Biol. 117:665-675. Planes, S., and G. Lecaillon. 1998. Consequences of the founder effect in the genetic structure of introduced island coral reef fish populations. Biol. J. Linn. Soc. 63:537-552. Pollock, D. D., D. J. Zwickl, J. A. McGuire, and D. M. Hillis. 2002. Increased taxon sampling is advantageous for phylogenetic inference. Syst. Biol. 51:664-671. Posada, D. 2001. The effect of branch length variation on the selection of models of molecular evolution. J. Mol. Evol. 52:434-444. Posada, D., and K. A. Crandall. 1998. MODELTEST: testing the model of DNA substitution. Bioinformatics 14:817-818. Quattro, J. M., and W. J. Jones. 1999. Amplification primers that target locus-specific introns in Actinopterygian fishes. Copeia 1999:191-196.
96 Randall, J. E. 1967. Food habits of reef fishes of the West Indies. Stud. Trop. Ocean 5:665-847. Rico, C., I. Rico, and G. Hewitt. 1996. 470 million years of conservation of microsatellite loci among fish species. Phil. Trans. Roy. Soc. B 263:549-557. Roberts, C. M., B. Halpern, S. R. Palumbi, and R. R. Warner. 2001. Designing marine networks: Why small, isolated protected areas are not enough. Con. Biol. In Pract. 2:10-17. Roberts, C. M., C. J. McClean, J. E. N. Veron, J. P. Hawkins, G. R. Allen, D. E. McAllister, C. G. Mittermeier, F. W. Schueler, M. Spalding, F. Wells, C. Vynne, and T. B. Werner. 2002. Marine biodiversity hotspots and conservation priorities for tropical reefs. Science 295:1280-1284. Robertson, D. R. 2001. Population maintenance among tropical reef-fishes: inferences from small island endemics. Proc. Natl. Acad. Sci. USA 98:5667-5670. Robins, C. R., G. C. Ray, and J. Douglas. 1986. A field guide to Atlantic coast fishes. Houghton Mifflin Co., New York, New York. Rocha, L. A., A. L. Bass, D. R. Robertson, and B. W. Bowen. 2002. Adult habitat preferences, larval dispersal, and the comparative phylogeography of three Atlantic surgeonfishes (Teleostei: Acanthuridae). Mol. Ecol. 11:243-252. Rocha, L. A., and I. L. Rosa. 1999. New species of Haemulon (Teleostei: Haemulidae) from northeastern Brazilian coast. Copeia 1999:447-452. Rocha, L. A., and I. L. Rosa. 2001. Baseline assessment of reef fish assemblages of Parcel Manuel Luiz Marine State Park, Maranho, north-east Brazil. J. Fish Biol. 58:985-998. Rocha-Olivares, A., N. M. Garber, and K. C. Stuck. 2000. High genetic diversity, large inter-oceanic divergence and historical demography of the striped mullet. J. Fish Biol. 57:1134-1149. Rogers, A. R. 1995. Genetic evidence for a Pleistocene population explosion. Evolution 49:608-615. Rogers, A. R., A. E. Fraley, M. J. Bamshad, W. S. Watkins, and L. B. Jorde. 1996. Mitochondrial mismatch analysis is insensitive to the mutational process. Mol. Biol. Evol. 13:895-902. Rogers, A. R., and H. Harpending. 1992. Population growth makes waves in the distribution of pairwise genetic distances. Mol. Biol. Evol. 9:552-569.
97 Rosen, D. E. 1976. A vicariance model of Caribbean biogeography. Syst. Zool. 24:431-464. Sadovy, Y. 2000. Regional survey of fry/fingerling supply and current practices for grouper mariculture: evaluating current status and long-term prospects for grouper mariculture in Southeast Asia. Collaborative APEC Grouper Research and Development Network, Hong Kong, China. Sadovy, Y. 2001. The threat of fishing to highly fecund fishes. J. Fish Biol. 59:S90-S108. Sadovy, Y., and A. S. Cornish. 2000. Reef fishes of Hong Kong. Hong Kong University Press, Hong Kong, China. Saiki, R. K., S. Scharf, F. Faloona, K. B. Mullis, G. T. Horn, H. A. Ehrlich, and N. Arnheim. 1985. Enzymatic amplification of beta-globin genomic sequences and restriction site analysis for diagnosis of sickle-cell anemia. Science 230:1350-1354. Santini, F., and R. Winterbottom. 2002. Historical biogeography of the Indo-western Pacific coral reef biota: is the Indonesian region a center of origin? J Biogeog 29:189-205. Schluter, D. 2000. The ecology of adaptive radiation. Oxford University Press, Oxford. Schneider, S., and L. Excoffier. 1999. Estimation of past demographic parameters from the distribution of pairwise differences when the mutation rates very among sites: Application to human mitochondrial DNA. Genetics 152:1079-1089. Schneider, S., D. Roessli, and L. Excoffier. 2000. Arlequin ver. 2.000: A software for population genetics data analysis. Genetics and Biometry Laboratory, Univ. of Geneva, Geneva, Switzerland. Shimodaira, H. 2002. An approximately unbiased test of phylogenetic tree selection. Syst. Biol. 51:492-501. Shimodaira, H., and M. Hasegawa. 1999. Multiple comparisons of log-likelihoods with applications to phylogenetic inference. Mol. Biol. Evol. 16:1114-1116. Shulman, M. J., and E. Bermingham. 1995. Early life histories, ocean currents, and the population genetics of Caribbean reef fishes. Evolution 49:897-910. Silva, J., W. A., M. C. R. Costa, V. Valente, J. de Freitas Sousa, M. L. Paco-Larson, E. M. Espreafico, S. S. Camargo, E. Monteiro, A. de Jesus Holanda, M. A. Zago, A. J. G. Simpson, and E. D. Neto. 2001. PCR template preparation for capillary DNA sequencing. BioTechniques 30:537-542.
98 Slowinski, J. B., and R. D. M. Page. 1999. How should species phylogenies be inferred from sequence data? Syst. Biol. 48:814-825. Smith, C. L. 1971. A revision of the American Groupers: Epinephelus and allied genera. Bull. Am. Mus. Nat. Hist. 146:241. Smith, J. L. B. 1964. A new serranid fish from deep water off Cook Island, Pacific. Ann. Mag. Nat. Hist. Ser. 13 6:719-720, plate 21. Smith, P. J. 1994. Genetic diversity of marine fisheries resources : possible impacts of fishing. Food and Agriculture Organization of the United Nations, Rome, Italy. Smith, P. J., R. I. C. C. Francis, and M. McVeagh. 1991. Loss of genetic diversity due to fishing pressure. Fish. Res. 10:309-316. Smith-Vaniz, W. F., G. D. Johnson, and J. E. Randall. 1988. Redescription of Gracila albomarginata (Fowler and Bean) and Cephalopholis polleni (Bleeker) with comments on the generic limits of selected Indo-Pacific groupers (Pisces: Serranidae: Epinephelinae). Proc. Acad. Nat. Sci. Philadelphia 140:1-23. Song, C. B. 1994. Molecular evolution of the cytochrome b gene among Percid fishes. Unpubl. diss., Univ. of Illinois at Urbana-Champaign, Urbana-Champaign, Illinois. Song, C. B., T. J. Near, and L. M. Page. 1998. Phylogenetic relations among percid fishes as inferred from mitochondrial cytochrome b. Mol. Phyl. Evol. 10:343-353. Stevens, P. F. 1992. Species: historical perspectives. Pp. 302-311 in: Keywords in evolutionary biology. E. F. Keller, and E. A. Lloyd (eds.). Harvard University Press, Cambridge, Massachusetts. Streelman, J. T., M. Alfaro, M. W. Westneat, D. R. Bellwood, and S. A. Karl. 2002. Evolutionary history of the parrotfishes: biogeography, ecomorphology, and comparative diversity. Evolution 56:961-971. Sweijd, N. A., R. C. K. Bowie, B. S. Evans, and A. L. Lopata. 2000. Molecular genetics and the management and conservation of marine organisms. Hydrobiologia 420:153-164. Swofford, D. L. 2002. PAUP* 4.0, Phylogenetic analysis using parsimony and other methods. Beta version 4.0b5. Sinauer Associates, Inc, Sunderland, Massachusetts.
99 Swofford, D. L., and D. P. Begle. 1993. PAUP: Phylogenetic analysis using parsimony, ver. 3.1. userâ€™s manual. Illinois Natural History Survey, Champaign, Illinois. Taberlet, P., A. Meyer, and J. Bouvet. 1992. Unusually large mitochondrial variation in populations of the blue tit, Parus caereleus. Mol. Ecol. 1:27-36. Takahashi, K., and M. Nei. 2000. Efficiencies of fast algorithms of phylogenetic inference under the criteria of maximum parsimony, minimum evolution, and maximum likelihood when a large number of sequences are used. J. Mol. Evol. 17:1251-1258. Tamura, K., and M. Nei. 1993. Estimation of the number of nucleotide substitutions in the control region of mitochondrial DNA in humans and chimpanzees. Mol. Biol. Evol. 10:512-526. Thresher, R. E. 1980. Reef fish: behavior and ecology on the reef and in the aquarium. Palmetto Publishing, St. Petersburg, Florida. Tringali, M. D., and R. R. Wilson Jr. 1993. Differences in haplotype frequencies of mtDNA of the Spanish sardine Sardinella aurita between specimens from the eastern Gulf of Mexico and southern Brazil. Fish. Bull. 91:362-370. Van Herwerden, L., C. R. Davies, and J. H. Choat. 2002. Phylogenetic and evolutionary perspectives of the Indo-Pacific grouper Plectropomus species on the Great Barrier Reef, Australia. J. Fish Biol. 60:1591-1596. Vecchione, M., M. F. Mickevich, K. Fauchald, B. B. Collette, A. B. Williams, T. A. Munroe, and R. E. Young. 2000. Importance of assessing taxonomic adequacy in determining fishing effects on marine biodiversity. ICES J. Mar. Sci. 57:677-681. Veron, J. E. N. 1995. Corals in space and time: The biogeography and evolution of the Scleractinia. Comstock/Cornell Publishing, Ithaca, New York, New York. Wakeley, J. 2000. The effects of subdivision on the genetic divergence of populations and species. Evolution 54:1092-1101. Ward, R. D., M. Woodward, and D. O. F. Skibinski 1994. A comparison of genetic diversity levels in marine, freshwater and anadromous fishes. J. Fish Biol. 44:213-232. Weaver, D. C. 1996. Feeding ecology and ecomorphology of three sea basses (Pisces: Serranidae) in the northeastern Gulf of Mexico. Unpubl. M.S. Thesis, Univ. of Florida, Gainesville, Florida. Wiens, J. J., and T. W. Reeder. 1995. Combining data sets with different numbers of taxa for phylogenetic analysis. Syst. Biol. 44:548.
100 Yang, Z. 1996. Among-site rate variation and its impact on phylogenetic analyses. TREE 11:367-372. Yoder, A. D., J. A. Irwin, and B. A. Payseur. 2001. Failure of the ILD to determine data combinability for slow loris phylogeny. Syst. Biol. 50:4084. Zane, L., L. Ostellari, L. Maccatrozzo, L. Bargelloni, B. Battaglia, and T. Patarnello. 1998. Molecular evidence for genetic subdivision of Antarctic krill (Euphausia superba Dana) populations. Proc. Roy. Soc. L. Ser. B 265:2387-2391. Zwickl, D. J., and D. M. Hillis. 2002. Increased taxon sampling greatly reduces phylogenetic error. Syst. Biol. 51:588-598.
BIOGRAPHICAL SKETCH Joel Carlin was born in Kalamazoo, Michigan in 1969; and was raised in Indiana. He earned a B.Sc. in marine biology from the University of North Carolina at Wilmington in 1991; and a M.Sc. in zoology from Louisiana State University in 1995. Since 1995 he has co-published six peer-reviewed articles and two reports on genetic variability in aquatic organisms and received ten competitive grants or awards. He currently resides in Gainesville, Florida and continues to develop as a teacher and researcher of marine science, evolution, and natural history. 101