Phylogeography of the Pantropical Sea Urchin Eucidaris in Relation to Land Barriers and Ocean Currents Author(s): H. A. Lessios, B. D. Kessing, D. R. Robertson and G. Paulay Reviewed work(s): Source: Evolution, Vol. 53, No. 3 (Jun., 1999), pp. 806-817 Published by: Society for the Study of Evolution Stable URL: http://www.jstor.org/stable/2640720 . Accessed: 17/01/2013 14:37 Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at . http://www.jstor.org/page/info/about/policies/terms.jsp . JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org. . Society for the Study of Evolution is collaborating with JSTOR to digitize, preserve and extend access to Evolution. http://www.jstor.org This content downloaded on Thu, 17 Jan 2013 14:37:51 PM All use subject to JSTOR Terms and Conditions Evolution, 53(3), 1999, pp. 806-817 PHYLOGEOGRAPHY OF THE PANTROPICAL SEA URCHIN EUCIDARIS IN RELATION TO LAND BARRIERS AND OCEAN CURRENTS H. A. LESSIOS,1,2 B. D. KESSING,1 D. R. ROBERTSON,1 AND G. PAULAY3 1Smithsonian Tropical Research Institute, Box 2072, Balboa, Panama 2E-mail. lessiosh @ naos. si. edit 3Marine Laboratory, University of Guam, Mangilao, Guam 96923, USA E-mail: gpaulay@ uog9. uog. edu Abstract.-The pantropical sea urchin genus Eucidaris contains four currently recognized species, all of them allopatric: E. metularia in the Indo-West Pacific, E. thouarsi n the eastern Pacific, E. tribuloides in both the western and eastern Atlantic, and E. clavata at the central Atlantic islands of Ascension and St. Helena. We sequenced a 640-bp region of the cytochrome oxidase I (COI) gene of mitochondrial DNA to determine whether this division of the genus into species was confirmed by molecular markers, to ascertain their phylogenetic relations, and to reconstruct the history of possible dispersal and vicariance events that led to present-day patterns of species distribution. We found that E. metularia split first from the rest of the extant species of the genus. If COI divergence is calibrated by the emergence of the Isthmus of Panama, the estimated date of the separation of the Indo-West Pacific species is 4.7-6.4 million years ago. This date suggests that the last available route of genetic contact between the Indo-Pacific and the rest of the tropics was from west to east through the Eastern Pacific Barrier, rather than through the Tethyan Sea or around the southern tip of Africa. The second cladogenic event was the separation of eastern Pacific and Atlantic populations by the Isthmus of Panama. Eucidaris at the outer eastern Pacific islands (Galapagos, Isla del Coco, Clipperton Atoll) belong to a separate clade, so distinct from mainland E. thouarsi as to suggest that this is a different species, for which the name E. galapagensis is revived from the older taxonomic literature. Complete lack of shared alleles in three allozyme loci between island and mainland populations support their separate specific status. Eucidaris gala- pagensis and E. thouarsi are estimated from their COI divergence to have split at about the same time that E. thouarsi and E. tribuloides were being separated by the Isthmus of Panama. Even though currents could easily convey larvae between the eastern Pacific islands and the American mainland, the two species do not appear to have invaded each other's ranges. Conversely, the central Atlantic E. clavata at St. Helena and Ascension is genetically similar to E. tribuloides from the American and African coasts. Populations on these islands are either genetically connected to the coasts of the Atlantic or have been colonized by extant mitochondrial DNA lineages of Eucidaris within the last 200,000 years. Although it is hard to explain how larvae can cross the entire width of the Atlantic within their competent lifetimes, COI sequences of Eucidaris from the west coast of Africa are very similar to those of E. tribuloides from the Caribbean. FST statistics indicate that gene flow between E. metularia from the Indian Ocean and from the western and central Pacific is restricted. Low gene flow is also evident between populations of E. clavata from Ascension and St. Helena. Rates of intraspecific exchange of genes in E. thouarsi, E. galapagenisis, and E. tribuloides, on the other hand, are high. The phylogeny of Eucidaris confirms Ernst Mayr's conclusions that major barriers to the dispersal of tropical echinoids have been the wide stretch of deep water between central and eastern Pacific, the cold water off the southwest coast of Africa, and the Isthmus of Panama. It also suggests that a colonization event in the eastern Pacific has led to speciation between mainland and island populations. Key words.-Biogeography, cytochrome oxidase, gene flow, islands, mitochondrial DNA, ocean currents, sea urchins. Received May 1, 1998. Accepted January 25, 1999. In 1954 Ernst Mayr published an article on geographic speciation of tropical echinoids (Mayr 1954). In this paper he plotted the ranges of species belonging to 16 sea urchin, sand dollar, and heart urchin genera as given in Mortensen's (1928-1951) monograph, and reached the conclusion that their geographic distributions were consistent with allopatric speciation. Geographical barriers seen by Mayr as causing speciation were the wide stretch of deep water dividing the eastern Pacific from the central Pacific (the Eastern Pacific Barrier), the Isthmus of Panama, and the cold waters off southwest Africa. He wrote that the genus Eucidatris "illus- trates geographic speciation almost diagrammatically." Mayr reached these conclusions while recognizing that he was dealing with morphospecies rather than biological spe- cies and that echinoid systematics were still at the stage of alpha taxonomy. We now have the tools to investigate what Avise et al. (1987) have called "phylogeography," that is, intra- and interspecific phylogenies, which in combination with distributional information can help deduce patterns and causes of divergence and speciation. Molecular phylogeog- raphy can provide information difficult to obtain from other characters, such as the existence of sibling species and the timing of splits between populations (Palumbi 1996, 1997). The present paper is an attempt to apply molecular tools to the species of the genus Eucidaris to evaluate Mayr's con- clusions. Eucidaris belongs to the subclass Perischoechinoidea, sep- arated from all other extant Echinoidea at least since the Triassic (Durham 1966; Smith 1984). The fossil record of Eucidaris dates back to the Upper Eocene (Fell 1966; Cutress 1980), approximately 50 million years ago (mya). The genus is pantropical, with all of its species concentrated in shallow water (< 570 m; Mortensen 1928-195 1). Its four recognized morphospecies have adjacent but nonoverlapping geograph- ical distributions. Eucidaris metularia ranges from the east coast of Africa to the central Pacific. Eucidaris thouarsi is found in the eastern Pacific. Eucidaris tribuloides is distrib- uted from the Atlantic coast of tropical America to the west- ern coast of Africa. Eucidaris clavata is endemic to the central Atlantic islands of Ascension and St. Helena. The morpho- 806 ?) 1999 The Society for the Study of Evolution. All rights reserved. This content downloaded on Thu, 17 Jan 2013 14:37:51 PM All use subject to JSTOR Terms and Conditions PHYLOGEOGRAPHY OF EUCIDARIS 807 logical differences between the species are slight. There are doubts in the taxonomic literature as to whether the Gala- pagos populations of Eucidaris belong to E. thouarsi or whether they constitute a separate species, E. galapagensis (Mortensen 1928-195 1, vol. I, p. 399) and as to whether E. clavata is a separate species from E. tribuloides (Mortensen 1928-1951, vol. I, p. 411). Eucidaris tribuloides on the Af- rican coast has been recognized as a separate subspecies, E. tribuloides africana (Mortensen 1928-1951, vol. I, p. 406). We used sequence data from a 640-bp fragment of the cytochrome oxidase I (COI) region of mitochondrial DNA (mtDNA) and isozymes to determine the validity of species described on the basis of morphology and to ascertain their phylogenetic relations. Specifically, we were interested in answering the following questions: (1) Is each geographic isolate a separate evolutionary lineage? (2) To what extent do the accepted morphospecies coincide with mtDNA clades? (3) What is the order and timing of cladogenic events, and what can they tell us about the effects of presumed geographic barriers on the speciation of tropical marine organisms? MATERIALS AND METHODS Collections One hundred twelve individuals were collected for mtDNA sequencing at the following locations for each species (see Fig. 1): E. metularia: Reunion, Indian Ocean (n = 4); Ishi- gaki-Jima, Sakishima Islands, Japan (n = 2); Sesoko, Oki- nawa-Jima, Ryukyu Islands, Japan (n = 4); Guam, west Pa- cific (n = 5); and Hawaii (n = 15, two specimens from Lanai, one from Maui, 12 from Oahu); E. thouarsi: Galapagos (n = 7, two specimens from each of the islands of Bartolome, Fernandina, and Genovesa, and one from Isabela); Isla del Coco (n = 6); Clipperton Atoll (n = 1); Guaymas, Sea of Cortez, Mexico (n = 6); and Taboguilla Island, Bay of Pan- ama (n = 4); E. tribuloides: Carrie Bow Cay, Belize (n = 11); Cochino Pequefno, Bay Islands, Honduras (n = 2); San Blas Islands, Panama (n = 10); Puerto Rico (n = 4); St. John, U.S. Virgin Islands (n = 3); Cayman Brac, Cayman Islands (n = 1); Tamandare, Brazil (n = 4); Sao Tome, Gulf of Guinea (n = 10); Ada, Ghana, Gulf of Guinea (n = 1); E. clavata: St. Helena (n = 7); Ascension (n = 5). We rooted the phy- logenetic tree of Eucidaris with one individual of Phylla- canthus imperialis from Rottnest Island, western Australia. Cladistic analysis of morphological characters suggests that Phyllacanthus is the sister genus of Eucidaris (Smith and Wright 1989). Samples were preserved in 95% ethanol, in high-salt DMSO buffer (Seutin et al. 1991), or in liquid N2. Forty-two individuals were collected for isozyme analysis at Isla del Coco (n = 20) and at Galapagos (n = 22, two in- dividuals from Isabela, five from Bartolome, four from Gen- ovesa, six from Marchena, two from Santiago, three from Fernandina); all were frozen in liquid N2. Mitochondrial DNA We amplified and sequenced 640 nucleotides from the COI region, corresponding to positions 6448 (or, depending on the primer, 6499) to 7128 of the mitochondrial genome of Strongylocentrotus purpuratus (Jacobs et al. 1988). Genomic DNA extractions, polymerase chain reaction (PCR) ampli- fication, PCR product purification, and DNA sequencing were carried out as described previously (Lessios et al. 1998). Primers for both PCR and sequencing were: either COI-f 5' (CCTGCAGGAGGAGGAGAYCC) or COI-p 5' (GGTCA- CCCAGAAGTGTACAT) and COI-a 5' (AGTATAAGC- GTCTGGGTAGTC). All individuals were sequenced in both directions. Phylogenies from mtDNA data were constructed with test version 4.0d64 of PAUP*, written by David L. Swof- ford and used with his permission, and with PUZZLE 4.0, written by Strimmer and von Haeseler (1996). Other analyses were performed with SEQUENCER 4.0, an Apple Macintosh program written by B. D. Kressing and available from him. Sequences have been deposited in GenBank under accession nos. AF063309-AF063399 and AF107700-AF107721. Isozymes Because the COI sequence data indicated that Eucidaris from the outer eastern Pacific islands (Galapagos, Isla del Coco, and Clipperton) were a distinct lineage from E. thouarsi from the American coast (see results), isozymes were used to determine whether genetic discontinuities in the eastern Pacific extended to the nuclear genome. Twelve loci were assayed in animals from the Galapagos and Isla del Coco in the same buffers and with the same staining recipes as those of Bermingham and Lessios (1993). The data were compared with data from the same loci obtained previously by these authors from coastal populations at Mexico and Panama. Pre- sumptive alleles were standardized by running individuals from Panama on the same gels as those from Galapagos and Isla del Coco. The assayed loci were: acid phosphatase (Acph), aspartate aminotransferase (Got-], Got-2), fructokin- ase (Fk), hexokinase (Hk), malate dehydrogenase (Mdh-]), L-leucyl-L-tyrosine peptidase (Peplt-]), phosphoglucose isomerase (Pgi), phosphoglucomutase (Pgm), octanol dehy- drogenase (Odh), superoxide dismutatse (To), and xanthine dehydrogenase (Xdh). Calculation and comparison of jack- knifed average Nei's (1978) genetic distances were performed according to the method of Mueller and Ayala (1982), em- ploying program NEIC (Lessios 1990). RESULTS Phylogenetic Groupings and Genetic Similarities Figure 2 presents a maximum-likelihood reconstruction of the phylogeny of Eucidaris, as inferred from the determined COI sequences. In addition to the presented tree produced by the quartet puzzling algorithm of Strimmer and von Hae- seler (1996), we also reconstructed phylogeny using the neighbor-joining algorithm of Saitou and Nei (1987), with distances corrected by Hasegawa et al.'s (1985) model of base substitution. The neighbor-joining tree was bootstrapped 1000 times. When branches with less than 70% support were collapsed, the tree topologies produced by the two techniques were identical, except for some minor rearrangements of the terminal branches. There is strong support for the existence of four clades (see Fig. 2). The Indo-Pacific E. metularia (clade A) is an outgroup to all other Eucidaris. We were able to obtain spec- This content downloaded on Thu, 17 Jan 2013 14:37:51 PM All use subject to JSTOR Terms and Conditions 808 H. A. LESSIOS ET AL. 1y ga~rJ O-_ Ad = o , UU 00 00:~~~~~~~~~~~~~~~~~~~~~~~~~~~~~0 NO C .-fnA I-! . - - - '0 * .. - f F * A~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~il E Sq > ? -4 S E 5~~~~~~~l4 -,I t j - 00 -4 (ND 7:'~~~~~~~~~~~~~~~~~~~~~~~~~~' l)0 rX- ; wil* D * ASH~~E~ S - - A' I O O s cd O z~~0)Q' This content downloaded on Thu, 17 Jan 2013 14:37:51 PM All use subject to JSTOR Terms and Conditions PHYLOGEOGRAPHY OF EUCIDARIS 809 Ascension-1, 2 81 Ascension-3; St. Helena-i E. clavata St. Helena-2, 3, 4, 5, 6, 7 St. John-I Puerto Rico- I San Blas-I San Blas-2 Honduras- I Belize- 1, 2; Caymans- 1; < Honduras-2; San Blas-3, 4, 5, 6; Belize-3 Puerto Rico-2; St. John-2 55% Belize-4 Belize-5 Belize-6 Belize-7; Puerto Rico-3 Belize-8 Brazil- 1 J Brazil-2 Puerto Rico-4; Sao Tom6-1; 74 Ghana Brazil-3, 4; San Blas-7, 8 E. tribuloides Sao Tom6-2 Sao Tom6-3 Sao Tom6-4 Sao Tome-5 Sao Tome-6 Sao Torn6-7 Sao Tom6-8 Sao Tom6-9 Belize-9 San Blas-9 Belize-10 St. John-3 94 Belize-i I Sao Tom6- IO San Blas-IO Coco- 1; Galapagos- I Galapagos-2 Galapagos-3 Galapagos-4 Coco-3 C - iCoco-32 E. galapagensis Coco-4 Coco-5, 6; Galapagos-5, 6, 7 Clipperton- I 88 Mexico-i 94 Mexico-2 Mexico-3; Panama-I ________________________ I B Mexico-4 E. thouarsi 100 Panama- 2,3 Panama-4 Mexico-5 Mexico-6 80 Hawaii-l Hawaii-2 - Reunion- 1 - Ishigaki-l Reunion-2 Reunion-3 Reunion-4 Hawaii-3 Hawaii-4 Hawaii-6 E. metularia Hawaii-7 Hawaii-8, 9 Hawaii- 10 Hawaii-lI, 12, 13, 14, 15 Sesoko-1 Ishigaki-2 Guam- 1, 2 Guam-3 Guam-4, 5; Sesoko-2, 3, 4 Phyllacanthus imperialis FIG. 2. Maximum-likelihood tree of COI haplotypes of Eucidaris, generated by the quartet puzzling technique of Strimmer and von Haeseler (1996), using Hasegawa et al. 's (1985) model of base substitution. Multiple labels next to a branch indicate that the same haplotype was observed in more than one individual. Numbers next to nodes indicate percent of quartets that support hem. Only nodes with more than 70% support are shown. Letters next to branches mark major clades. This content downloaded on Thu, 17 Jan 2013 14:37:51 PM All use subject to JSTOR Terms and Conditions 810 H. A. LESSIOS ET AL. TABLE 1. Number of sampled individuals and gene frequencies in 12 electrophoretically assayed loci of eastern Pacific populations of Eucidaris. Locus abbreviations are explained in the text. Allele designations indicate relative distance from the origin. Data for E. thouarsi are from Bermingham and Lessios (1993). Got-2, Odh, and Xdh were monomorphic in all populations. E. galapagensis E. thouarsi _____________________ Isla del Locus Allele Mexico Taboguilla Coco Galapagos Acph n 30 28 18 22 90 0.133 0.196 0.472 0.546 100 0.833 0.786 0.528 0.455 110 0.033 0.018 0.000 0.000 Fk n 19 33 20 22 80 0.000 0.015 0.000 0.000 90 1.000 0.924 0.275 0.273 98 0.000 0.000 0.650 0.568 100 0.000 0.061 0.000 0.000 120 0.000 0.000 0.075 0.159 Got-i n 32 76 16 22 100 1.000 1.000 0.000 0.000 110 0.000 0.000 1.000 1.000 Hk n 29 44 20 22 90 0.000 0.045 0.000 0.000 100 1.000 0.955 0.000 0.000 110 0.000 0.000 1.000 1.000 Mdh-i n1 37 55 19 22 97 0.000 0.000 0.842 0.932 98 0.000 0.009 0.000 0.000 100 0.405 0.109 0.158 0.068 103 0.595 0.855 0.000 0.000 105 0.000 0.027 0.000 0.000 Peplt-] n 33 56 20 22 80 0.879 0.000 0.000 0.000 90 0.121 0.170 0.100 0.114 100 0.000 0.750 0.900 0.886 110 0.000 0.080 0.000 0.000 Pgi n 30 26 19 94 0.000 0.077 0.000 95 0.000 0.077 0.105 97 0.050 0.000 0.000 100 0.900 0.692 0.711 103 0.050 0.154 0.184 Pgm n 25 35 18 22 70 0.020 0.057 0.000 0.000 90 0.100 0.157 0.028 0.023 100 0.640 0.671 0.944 0.977 110 0.200 0.071 0.028 0.000 120 0.040 0.043 0.000 0.000 To , 22 46 20 22 100 0.000 0.000 1.000 1.000 105 1.000 1.000 0.000 0.000 imens of this species only from five localities, and with this limited sample it is difficult to draw conclusions regarding the number of distinct lineages that are truly represented in this vast oceanic expanse. However, despite the long distance separating Reunion from Guam (a straight-line distance of > 10,000 km) and from Hawaii (17,000 km), the mtDNA lineages do not form separate clades, indicating that even very distant populations of E. metularia have not evolved independently. The data clearly indicate that there are two distinct lineages of Eucidaris in the eastern Pacific. All coastal specimens from Mexico and Panama belong to one clade (B), which repre- TABLE 2. Average jackknifed means of Nei's D distances between eastern Pacific populations, assayed for 12 electrophoretically de- termined loci. Mexico Taboguilla Isla del Coco Taboguilla 0.084 Isla del Coco 0.639 0.580 Galapagos 0.718 0.625 0.000 sents E. thouarsi sensu stricto. Specimens from Galapagos, Isla del Coco and the Clipperton Atoll cluster in a separate clade (C), which is likely to be a different species. Reviving D6derlein's (1887) name, we have labeled it as E. galapa- gensis. The 640 mtDNA bases we sampled contain 33 nu- cleotide sites in which E. thouarsi and E. galapagensis are consistently different. In keeping with the high degree of conservatism of amino acid composition of COI, all of these fixed differences are silent. The single specimen we were able to obtain from Clipperton Atoll is sufficiently different from all others coming from eastern Pacific islands (Kimura [1980] two-parameter corrected percent difference [K2] of 1.79%) to suggest that it may belong to an additional evo- lutionary lineage. The differentiation of Eucidaris at this re- mote Atoll from other populations of the eastern Pacific needs to be investigated further. Isozyme results from eastern Pacific populations confirm that Eucidaris populations from Isla del Coco and from the Galapagos are very similar to each other and that no gene flow has occurred between island and mainland populations during recent evolutionary time. The most common allele in five of the 12 assayed loci is different between the islands and the American coast. In three of these loci there is com- plete partitioning of alleles (Table 1). Nei's D between pop- ulations at Galapagos and Isla del Coco is less than 0.001 (Table 2), whereas the average jackknifed Nei's D between E. galapagensis and E. thouarsi is 0.641, which is not sig- nificantly different (P > 0.05) from the value of Nei's D (0.758) calculated from the same 12 loci (of 25 sampled by Bermingham and Lessios 1993) between E. thouarsi and the Caribbean E. tribuloides. Clade D in Figure 2 is composed of all individuals from the Atlantic, irrespective of whether they belong to the west Atlantic E. tribuloides, the African E. tribuloides africana, or the central Atlantic E. clavata. Within this clade, there are no resolved subclades, except for one joining all haplotypes from Ascension and one from St. Helena (which is identical to one of the Ascension haplotypes). The low degree of phy- logenetic structure is the result of high haplotype similarity. There are no fixed differences between Caribbean, Brazilian, and African E. tribuloides. A single transition at a third codon position is diagnostic between E. clavata from Ascension (and one individual from St. Helena) and E. tribuloides from the American and African coasts. A third position transition at a different site distinguishes the remaining six individuals from St. Helena from continental E. tribuloides. Thus, some differentiation exists between E. clavata and E. tribuloides. However, the mtDNA divergence between Eucidaris from the Caribbean and the central Atlantic islands (average K2 = 0.49) is no larger than the differentiation seen between E. clavata from St. Helena and from Ascension (K2 = 0.64). This content downloaded on Thu, 17 Jan 2013 14:37:51 PM All use subject to JSTOR Terms and Conditions PHYLOGEOGRAPHY OF EUCIDARIS 811 TABLE 3. FST-values (below the diagonal) and estimated number of female propagules (Nen; above the diagonal) between populations of Eucidaris in which at least four individuals were sampled. Calculations follow Hudson et al. (1992); ud, undefined value of Nem because FST for this comparison is negative. Indo-West Pacific Species Region Reunion Ryukyu Guam Hawaii E. metularia Reunion 0.79 0.53 0.22 E. metularia Ryukyu 0.39 ud 0.88 E. metularia Guam 0.48 -0.09 1.23 E. metularia Hawaii 0.70 0.36 0.29 Eastern Pacific Species Region Panama Mexico Isla del Coco Galapagos E. thouarsi Panama ud 0.02 0.02 E. thouarsi Mexico -0.03 0.04 0.04 E. galapagensis Isla del Coco 0.96 0.93 ud E. galapagensis Galapagos 0.95 0.93 -0.06 Atlantic Species Region Ascension St. Helena Puerto Rico Belize San Blas Brazil Sao Torn6 E. clavata Ascension 0.19 0.51 1.44 0.66 0.55 1.11 E. clavata St. Helena 0.73 0.53 6.28 1.12 0.52 4.77 E. tribuloides Puerto Rico 0.50 0.49 ud ud ud ud E. tribuloides Belize 0.26 0.07 -0.36 ud ud 4.50 E. tribuloides San Blas 0.43 0.31 -0.10 -0.07 596.41 2.31 E. tribuloides Brazil 0.48 0.49 -0.08 -0.22 0.01 ud E. tribuloides Sao Tome 0.31 0.09 -0.16 0.10 0.18 -0.11 Similarly, the difference between E. clavata and E. tribuloides africana (K2 = 0.75) is only slightly larger than the differ- ences within E. clavata. Estimates of gene flow within the Indo-West Pacific based on FST-values (Table 3) suggest that, despite the lack of dis- tinct evolutionary lineages, some degree of genetic isolation exists between the Indian and the Pacific Oceans. Although populations from Guam and the Ryukyu Islands are panmic- tic, and Hawaii and Guam are also connected by fairly high gene flow, Reunion, as might be expected, is more differ- entiated. Gene flow between this Indian Ocean island and the western Pacific is less than the equivalent of one female propagule per generation. Overall, the values of FST in E. metularia are correlated with the straight-line distance be- tween localities. In the eastern Pacific, mtDNA differentiation between localities within each presumed species is less than diversity within each locality. However, estimated migration between mainland E. thouarsi and island E. galapagensis is almost equal to zero, confirming our conclusion that island and mainland populations must have remained isolated for a long time. In the Atlantic, most populations are genetically connected. The two populations in the Caribbean from which we sampled 10 or more individuals, San Blas and Belize, contain hap- lotypes that are more similar between than within localities, leading to a negative FST-value. Eucidaris tribuloides from Brazil, despite substantial distance (4000 km along the coast from the closest point in the Caribbean) and possible barriers to echinoid dispersal due to the freshwater discharge of the Orinoco and the Amazon, is also connected to the Caribbean with high levels of gene flow. Indeed, the FST-values suggest that Eucidaris from the entire western Atlantic belong to a single panmictic unit. The overall FST-value for comparisons among all Caribbean localities plus Brazil (-0.143) is not only negative, but also not significantly different (P = 0.938) from that generated from 500 random reshufflings of hap- lotypes. There is also high apparent gene flow between the American and the African coast. Each western Atlantic pop- ulation of E. tribuloides exchanges genes with Sao Tome at rates higher than 2.3 females per generation (Table 3). How- ever, the genetic connections between E. tribuloides and E. clavata seem more variable. High gene flow is estimated between the African coast and the central Atlantic islands and between the central Atlantic islands and Belize. Puerto Rico, San Blas, and Brazil, despite being genetically con- nected with Belize and with the African coast, appear to be contributing fewer genes toward connections with E. clavata. Interestingly, although Ascension and St. Helena are sup- posed to be both inhabited by E. clavata, and even though the distance between them (1300 km) is smaller than that between the American and the African coast (a minimum of about 2800 km), the estimated number of female propagules exchanged between the two islands is the lowest for any comparison within the Atlantic. Timing of Cladogenic Events How long ago did these clades diverge from each other? Mitochondrial DNA differentiation (see Table 4), under the assumption of constant rates of evolution, can provide an estimate. This estimate cannot be very precise, because in Eucidaris evolution of the sequenced segment is not all that constant. If there were no rate variation between lineages, E. metularia would have been equidistant from all other species and E. tribuloides would be equidistant from E. galapagensis and E. thouarsi. However, the average K2 distance between E. metularia and each of the other species ranges from 16.01% (E. rnetularia-E. clavata) to 19.85% (E. rnetularia-E. gala- This content downloaded on Thu, 17 Jan 2013 14:37:51 PM All use subject to JSTOR Terms and Conditions 812 H. A. LESSIOS ET AL. TABLE 4. Means of nucleotide percent difference between species of Eucidaris. Values below the diagonal are percent difference in all sites (K9), corrected with Kimura's (1980) two-parameter model. Values on the diagonal (in bold) are average K9-values between individuals within a species. Values above the diagonal are percent nucleotide differences in silent sites (Ks), estimated from the equations of Pamilo and Bianchi (1993) and Li (1993). Species Region E. clavata E. meti/aria E. galapagensis E. thouarsi E. tribidloides E. clavata Central Atlantic 0.42 72.72 28.42 39.30 1.53 E. metularia Indo-West Pacific 16.01 1.12 131.74 90.52 74.55 E. galapagensis Central eastern Pacific 8.29 19.85 0.58 31.12 29.21 E. thouarsi East eastern Pacific 10.83 18.78 9.11 0.72 40.28 E. tribuloides Caribbean and African coast 0.54 16.34 8.49 11.05 0.52 pagensis), and the average difference of E. tribuloides from E. thouarsi (11.05 %) is 1.3 times higher than the difference of E. tribuloides from E. galapagensis (8.49%). Thus, K2- values that would have been equal under a strict molecular clock hypothesis can vary by up to 30%. Substitution at silent sites is supposed to be more regular because of the absence of selective constraints, yet Ks-values show variation that is even higher, probably because they are based on a fraction of all sites, and thus have larger sampling error (Table 4). Thus, the cladogenic events in Eucidaris can only be dated within broad time constraints. If we assume that the sepa- ration of E. tribuloides from E. thouarsi and E. galapagensis occurred at the time of the final closure of the Isthmus of Panama 3.1 million years ago (Coates and Obando 1996) and equate this length of time with the average divergence be- tween all east Pacific and all Atlantic sequences (K2 = 9.52%), we obtain a calibration of 3.1 % sequence divergence per million years. By this calibration, and allowing for a range of rate variation 15% on either side of the mean, E. metularia diverged from all other Eucidaris in the late Miocene/early Pliocene between 4.7 and 6.4 million years ago (average K2 = 17.19%), and E. galapagensis separated from E. thouarsi in the Pliocene approximately 2.5-3.4 million years ago. If we were to assume that E. clavata no longer exchanges genes with E. tribuloides, the estimated date of separation would be 148,000-200,000 years ago. DISCUSSION Correspondence between ntDNA Clades and Biological Species What new insights into the phylogeny of Eucidaris do the molecular data provide, and what general patterns of ocean biogeography can they address? The first point genetic data can establish, a point difficult to examine with previously available alpha taxonomic studies of the genus, is to dem- onstrate which populations exchange genes and which do not. In its extreme form of complete cessation of gene flow, this is tantamount to the demonstration of specific status. As mentioned in the introduction, there has been disagree- ment as to whether the eastern Pacific harbors one or two species of Eucidaris. Although the name E. thonarsi is un- questionably applied by all authors to the populations on the western American coast, D6derlein (1887) described the form found in the Galapagos as a separate species, E. galapagensis. However, Clark (1925) stated that he would hesitate to rec- ognize it even as a "valid variety." Mortensen (1928-1951) did accept E. galapagensis as a variety of E. thouarsi, but doubted that it deserved specific rank. More recent biogeo- graphic (Maluf 1988, 1991) and ecological (Glynn et al. 1979; Glynn 1994; Wellington 1997) works have all dealt with Galapagos populations of Eucidaris as if they belonged to E. thouarsi. However, both our mtDNA and our isozyme data provide strong evidence that E. galapagensis is a species separate from E. thouarsi, from which it diverged at about the same time that the eastern Pacific was being separated from the Caribbean by the Central American isthmus. More- over, this species is not a Galapagos endemic, but is also found at Isla del Coco and perhaps at Clipperton Atoll. An interesting question is whether E. galapagensis and E. thouarsi are truly allopatric. Given their morphological sim- ilarity, it is possible that colonization of the mainland by E. galapagensis or of the islands by E. thouarsi could have gone undetected in faunal lists of either region. Indeed, based on morphology, we have erroneously reported E. thouarsi as present at Isla del Coco and the Clipperton Atoll (see table 1 in Lessios et al. 1996). Although the question cannot be answered with certainty until the western American coast is more extensively surveyed, our collections would be unlikely to have missed genotypes of the "wrong" species in localities that were sampled. Our combined mtDNA and isozyme sam- ples come to 45 individuals from the islands (some individ- uals were assayed for both kinds of molecules) and 117 from the American mainland. Fifty-two additional E. thouarsi from the Bay of Panama and 51 from the Gulf of Chiriqui were sampled in previous isozyme studies (Lessios 1979, 1981) that included four of the loci diagnostic between E. thouarsi and E. galapagensis. These studies would have noted the presence of E. galapagensis genotypes in the mainland, yet none of them showed evidence of invasion of one lineage into the geographical range of the other. That no island ge- notypes appear to be present on the American coast is not easy to explain, given the relatively small distances involved (Isla del Coco is 500 km west of Costa Rica) and the pre- vailing current flow from west to east (Wyrtki 1965, 1966, 1967; Abbott 1966; Tsuchiya 1974), but neither is it all that unusual. There are 17 echinoderm species endemic to the Galapagos and nine endemic to Isla del Coco (Maluf 1991). The apparent absence of the reverse colonization, by E. thouarsi from the coast to the islands, is more peculiar. All other species of sea urchins found commonly on the mainland are also known from the outer islands. In at least one case, that of Diadenia nmexicanumn, sequencing of mtDNA has shown that this is not another case of an undetected sibling species (Lessios et al. 1996). Despite the geographic analogies between the eastern Pa- This content downloaded on Thu, 17 Jan 2013 14:37:51 PM All use subject to JSTOR Terms and Conditions PHYLOGEOGRAPHY OF EUCIDARIS 813 cific islands of Galapagos, Coco, and Clipperton and the cen- tral Atlantic islands of St. Helena and Ascension, the case of E. clavata is opposite to that of E. galapagensis. As they have for E. galapagensis, echinoderm taxonomists have ex- pressed doubts as to whether E. clavata should be treated as a species distinct from E. tribuloides. Mortensen (1928-1951, vol. I, p. 411) considered whether E. clavata should be just a variety of E. tribuloides, but decided to describe it as a separate species and continued to regard it as such in other publications (Mortensen 1932, 1933). However, he stated that E. tribuloides is present at Ascension along with E. clavata (Mortensen 1936). Pawson (1978) carried out measurements of specimens of Eucidaris from Ascension, St. Helena, West Africa, and the West Indies, and, on the basis of morpho- logical differences, he concluded that E. clavata is a separate species to which all Eucidaris from both Ascension and St. Helena belong. In contrast to the case of E. galapagensis, however, the mtDNA similarities indicate that E. clavata is more likely to be a peripheral deme of E. tribuloides. Although the spatial distances are formidable (Ascension to Brazil, 2300 km, Ascension to St. Helena, 1300 km, St. Helena to Africa, 1900 km), transport of larvae from the continents to the central Atlantic is not beyond the realm of possibility. The larvae of E. thouarsi (Emlet 1988) and (pos- sibly) E. tribuloides (McPherson 1968) settle in the laboratory 25-30 days after fertilization. The southernmost limit of the range of E. tribuloides extends to Rio de Janeiro (Bernasconi 1955) and the easternmost limit to the Gulf of Guinea (Chesh- er 1966). The close genetic similarity between E. tribuloides from the Caribbean Sea and Brazil to the Gulf of Guinea indicates that there has been recent gene flow between these areas, despite the fact that the transit time across the tropical Atlantic, estimated from the average velocity of the equatorial undercurrent (Metcalf et al. 1962), is 43-70 days (Chesher 1966). The time estimated by Manning and Chace (1990) for larval transit from the coast of Brazil to Ascension on the equatorial counter-current is 48 days. Travel in the opposite direction, from the central Atlantic to the coast of Brazil on the westward-flowing south equatorial current (Longhurst 1962) is also a possibility. Edwards and Lubbock (1983) report E. clavata from St. Paul's Rocks, which is 2000 km northwest of Ascension, and 960 km northeast of the Bra- zilian coast, so there may be a stepping stone along the way. That FST-values suggest little genetic exchange between the two central Atlantic islands is partly due to the low mtDNA variability within each locality (see below), which tends to drive FST values upward (Charlesworth 1998). However, as pointed out, mtDNA sequence divergence between haplo- types from the two islands is larger than differentiation be- tween central and western Atlantic populations of Euicidaris. As the single shared haplotype indicates, migration between St. Helena and Ascension does occasionally happen. How- ever, each island represents a minuscule source and a mi- nuscule target for waterborne larvae, so it is not surprising that such events are infrequent. The alternative explanation to ongoing gene flow between the central Atlantic islands and the continental margins is that single, improbable events of colonization established the ancestors of the extant mtDNA haplotypes at each island and there has been little or no subsequent genetic exchange. St. TABLE 5. Sample size and average within-population percent nu- cleotide difference (K9) of COI in all populations of Eucidaris in which more than two individuals were sampled. Locality nI K2 Reunion 4 2.33 Ryukyu 4 0.25 Guam 5 0.31 Hawaii 15 0.63 Galapagos 7 0.44 Isla del Coco 6 0.33 Bay of Panama 4 0.36 Mexico 6 0.89 Belize 11 0.54 San Blas 10 0.31 St. John 3 0.85 Puerto Rico 4 0.28 Brazil 4 0.34 Sao Tome 10 0.72 St. Helena 7 0.14 Ascension 5 0.25 Helena is thought to be 13.3-15.3 million years old, but Ascension has only emerged in the last 1.5 million years (Baker 1970; Mitchell-Thome 1982). If we assume that there has not been repeated genetic contact, our calibration of the Eucidaris COI clock would estimate the colonization of these islands as having occurred less than 200,000 years ago. Such a recent date eliminates the possibility that colonization by extant mtDNA clades in the central Atlantic took place at a time before sea-floor spreading had established the Atlantic in its present-day dimensions (Berggren and Hollister 1974; Scheltema 1995). Haplotype diversity of Eucidaris in the central Atlantic islands suggests that a single recent introduction is a pos- sibility. Palumbi et al. (1997) found that Echinonietra mathaei in Hawaii shows 0% sequence heterogeneity in COI. They explained this lack of diversity as the result of a single recent colonization event. Eucidaris at St. Helena has less haplotype diversity than any other population of this genus for which we sequenced three or more specimens (see Table 5). Six haplotypes are identical, while the seventh is similar to the ones from Ascension (see Fig. 2). Diversity values at As- cension are higher, but still among the lowest ones observed. Thus, although Eucidaris in the central Atlantic islands is not as genetically depauperate as E. mathaei at Hawaii and its haplotypes are closely related to those found at continental margins, it may be to a considerable extent cut off from the mainland. Possible Causes of Isolation Although Mayr (1954) was more inclined to think in terms of dispersal, the allopatric species of Eucidaris, with adjacent ranges spanning in their aggregate the entire tropics, have always been consistent with the hypothesis of a circumglobal original common stock, which became fractionated by the closure of seaways and the formation of deep oceanic stretch- es. The phylogenetic information added by our study cannot distinguish between biogeographic models of dispersal and vicariance, but it can provide estimates of the approximate age of separation between geographical isolates. This, in turn, This content downloaded on Thu, 17 Jan 2013 14:37:51 PM All use subject to JSTOR Terms and Conditions 814 H. A. LESSIOS ET AL. allows informed guesses about the nature of the possible barriers that resulted in the species as we see them today. The first vicariance event in the history of extant Eucidaris was the isolation of the Indo-Pacific species from all others. In this regard, the phylogeny of Eucidaris resembles that of Echinometra, which also shows a deep split between species from the western Pacific and those from the eastern Pacific and the Atlantic (Palumbi 1996). Cutress (1980, p. 165) sug- gested that affinities between Atlantic and Indo-Pacific ci- daroids can be best explained as the result of west-to-east dispersal through the Tethyan Sea. Miocene fossils of Eu- cidaris from the Azores and the Canary Islands (areas in which the genus is now extinct) support the view of unin- terrupted distribution from the Atlantic to the Indian Ocean. However, our estimated time of separation of E. nmetularia suggests that circumglobal genetic connections in Eucidaris continued after the closure of the Tethyan Sea. Although there is considerable disagreement regarding the time of the final closure of the Tethys (Adams 1967; McKenzie 1967; Rug- gieri 1967; Luyendyk et al. 1972; Berggren and Hollister 1974; R6gl and Steininger 1984; Piccoli et al. 1987; Robba 1987; Rosen and Smith 1988; Por 1989; Hallam 1994, p. 178), we are aware of no estimate that is more recent than 12 million years ago; the most widely accepted date is the early Miocene, approximately 20 million years ago (Rosen 1984; Winterbottom and McLennan 1993). Even with broad allowances for rate variation, our oldest estimate from mtDNA for the separation of E. metularia is 6.5 million years ago, in the late Miocene. Thus, Atlantic and eastern Pacific Eucidaris were still exchanging genes with the Indo-Pacific after the shallow epicontinental Tethyan Sea ceased to exist. In principle, connections were still possible in the late Mio- cene and early Pliocene around the southern tip of Africa, because the Benguela cold water upwelling system off south- western Africa, which appeared in the Miocene (Diester- Haass and Schrader 1979; Siesser 1980), was not established as a continuous phenomenon until the late Pliocene (Shannon 1985). However, unlike other Indo-Pacific shallow-water echinoids, the ranges of which extend past Durban, E. nme- tularia only reaches Mozambique (Clark 1925; Clark and Courtman-Stock 1976). This suggests that Eucidaris may be a more "tropical" genus, the larvae of which are less likely to cross an area of even weak upwelling. A more probable last obstacle to migration is the Eastern Pacific Barrier, the 5000-km stretch of deep water separating the central from the eastern Pacific (Ekman 1953; Briggs 1974; Vermeij 1978, p. 253, 1987; Grigg and Hey 1992). Although the Eastern Pacific Barrier has possibly existed for the entire Cenozoic (Grigg and Hey 1992), in sea urchins, as in many other groups, it is functioning to the present day as a haphazard filter, which is subject to sporadic breaches by larvae (Lessios et al. 1996), sometimes resulting in massive gene flow (Les- sios et al. 1998). The presence in the eastern Pacific of mi- tochondrial haplotypes that belong to central Pacific species has been documented in other sea urchin genera, such as Echinothrix (Lessios et al. 1998), Diaderna (Lessios et al. 1996), and Echinonietra (Palumbi 1997). However, that cen- tral and eastern Pacific populations of some of these genera have speciated in the first place indicates that there have also been periods during which they did not maintain genetic con- tact. Apparently the last time that east and west Pacific Eu- cidaris exchanged genes was 4.7-6.4 million years ago. The cause of the next cladogenic event in the history of extant species of Eucidaris is no mystery. The Isthmus of Panama separated E. tribuloides from eastern Pacific stock. We cannot be certain whether this separation of Eucidaris was contemporaneous with the final closing of the isthmus or whether it occurred at some point during its gradual shoal- ing (Coates and Obando 1996) because the transisthmian COI divergence reported here (9.5% sequence divergence when distances between E. tribuloides and the two east Pacific spe- cies are averaged) is similar to that found between analo- gously separated Atlantic and Pacific species of Echinometra (9.8%), but twice as large as that of species of Diadema (4.7%; Lessios 1998). Although it is unclear whether the lower values of transisthmian divergence in Diaderna have resulted from slower rate of COI evolution in this genus, breaching of the isthmus at times of high sea level stands (Cronin and Dowsett 1996), continuing gene flow through mangrove swamps at the time of isthmus completion (Keller et al. 1989), or circumglobal gene flow around the southern tip of Africa, the discrepancy points to the possibility that separation of eastern Pacific and western Atlantic Eucidaris may be somewhat older than 3.1 million years ago. If so, the dates of other cladogenic events, dated by a calibration based on this split, may also be older, but not by so much as to alter conclusions as to the possible causes of separation. One would need to postulate a calibration of COI evolution three times slower than our estimate to accept the closure of the Tethys as a credible cause of speciation of E. nmetularia. The final definite split in Eucidaris occurred in the eastern Pacific between mainland and island populations. There is no obvious extrinsic barrier to gene flow that could have caused this split. Presumably, the shoaling of the Isthmus of Panama affected ocean circulation (Maier-Reimer et al. 1990; Haug and Tiedemann 1998), but it is not clear how it did so in the eastern Pacific or how long before the final closure the shoaling altered surface currents. It appears that, like many other marine species that are endemic to the Galapagos (Walker 1966; Glynn and Wellington 1983; James 1984; Garth 1991; Kay 1991; Maluf 1991; Allen and Robertson 1994) and/or to other eastern Pacific islands (Hertlein 1963; Garth 1991; Maluf 1991; Allen and Robertson 1994; Rob- ertson and Allen 1996), E. galapagensis may have speciated because larvae reached the islands and then, for unknown reasons, were cut off from the mainland. That the age of the oldest present-day Galapagos islands (see Glynn and Wel- lington 1983, p. 162; Chavez and Brusca 1991 and references therein) roughly agrees with the three million years ago tim- ing of the split between E. thouarsi and E. galapagensis is probably a coincidence. The central eastern Pacific contains several "drowned" islands, dating back to nine million years ago (Christie et al. 1992). Land vertebrates on the Galapagos appear to have been separated from their mainland ancestors for much longer than three million years (Wyles and Sarich 1983; Rassmann 1997). Thus, the time available for peripatric speciation of E. galapagensis is longer than the age of any extant island. This content downloaded on Thu, 17 Jan 2013 14:37:51 PM All use subject to JSTOR Terms and Conditions PHYLOGEOGRAPHY OF EUCIDARIS 815 Conclusions Because of the lack of geographic overlap between the species, the genus Eucidaris represents the simplest possible case for reconstructing phylogeographic patterns in pantrop- ical shallow water tropical genera with moderately long-lived larvae. Most of what we know about marine biogeography and barriers to dispersal for marine organisms has been based on the traditional approach of studying the number of species that are found in common in different geographic regions (e.g., Ekman 1953; Briggs 1974; Vermeij 1978, 1987). Sev- eral features of the Eucidaris phylogeography presented here have significance that may extend beyond this particular ge- nus. Despite the tremendous distances involved, populations from the east and west Atlantic coasts are connected by recent gene flow, as one would have surmised from their assignment by traditional taxonomists into the same species. The central Atlantic islands of Ascension and St. Helena show only a small degree of genetic isolation from the continental mar- gins. Estimates of the isolation of these islands, based on the percentage of endemism (e.g., Mortensen 1933; Briggs 1974; Rosewater 1975; Pawson 1978; Lubbock 1980; Edwards 1990; Manning and Chace 1990; Biernbaum 1996), may in some cases be biased upward because of the erroneous ele- vation of their populations to separate specific status. The offshore islands of the eastern Pacific have provided oppor- tunities for speciation dating back to the Pliocene. Despite the existence of oceanographic conditions favoring dispersal between them and the American mainland, the island and the mainland species have not invaded each other's ranges. This isolation is not limited to the Galapagos, as suggested by traditional biogeography (Ekman 1953, p. 45; Briggs 1974, pp. 43,50), but also extends to Isla del Coco, which has been included by some biogeographers (e.g., Hertlein 1963; Briggs 1974, p. 53) in the Panamic Province. As has been suggested numerous times on the basis of many groups of animals (for review, see Lessios 1998), the Isthmus of Panama represents a distinct biogeographic barrier that closed a primary avenue of dispersal between the tropical Atlantic and tropical Pacific. The last available route of genetic contact between the Indo- Pacific and the rest of the tropics was probably from west to east through the Eastern Pacific Barrier, rather than through the Tethys or around the southern tip of Africa. As more genera of marine organisms are examined in the same manner as we have done for Eucidaris, there will un- doubtedly be many that do not conform to one or more of these patterns. There may be some that conform to none of them, not even those that have once been thought so obvious as to need no study, such as the assumption that there has been no migration between the Indian Ocean and the tropical Atlantic after the closure of the Isthmus of Panama. The time- honored approach of comparing patterns found in many groups of organisms will determine their generality. ACKNOWLEDGMENTS We thank S. Williams for specimens from Brazil, H. Ban- ford for the specimen from Ghana, and P. Barnes for Phyl- lacanthus from western Australia. J. Eberhard, M. McCart- ney, J. Pearse, and S. Williams commented on the manuscript. A. Calder6n, L. Calder6n, C. Luna, and M. Soriano helped in the laboratory. The Administrator of Ascension, the Gov- ernment of St. Helena, H. Pinto da Costa, and the Direcao das Pescas of Sao Tome made collections in the central and eastern Atlantic possible. The captains and the crews of the R/V Benjamin, the M/S Marigold, the R/V Gyre, and the M/S Royal Star assisted with collections in the eastern Pacific. This study was supported by funds from the Smithsonian Molecular Evolution Program, by Smithsonian Scholarly Studies Grant 1234S607, Baird Fund Grant 14404507, and a Japan Society for the Promotion of Science fellowship to DRR. LITERATURE CITED ABBOTT, D. P. 1966. Factors influencing the zoogeographic affin- ities of the Galapagos. Pp. 108-112 in R. I. Bowman, ed. The Galapagos. Univ. of California Press, Berkeley. ADAMS, C. G. 1967. Tertiary Foraminifera in the Tethyan, Amer- ican, and Indo-Pacific provinces. Syst. Assoc. Publ. no. 7:195- 217. ALLEN, G. R., AND D. R. ROBERTSON. 1994. Fishes of the tropical eastern Pacific. Univ. of Hawaii Press, Honolulu. AVISE, J. C., J. ARNOLD, R. M. BALL, E. BERMINGHAM, T. LAMB, J. E. NEIGEL, C. A. REEB, AND N. C. SAUNDERS. 1987. Intra- specific phylogeography: the mitochondrial DNA bridge be- tween population genetics and systematics. Annu. Rev. Ecol. Syst. 18:489-522. BAKER, I. 1970. Geological history of Saint Helena in relation to its floral and faunal colonization. Mus. R. Afr. Cent. Tervuren Belg. Ann. Ser. 8 Sci. Zool. 181:25-35. BERGGREN, W. A., AND C. D. HOLLISTER. 1974. Paleogeography, paleobiogeography and the history of circulation in the Atlantic Ocean. Pp. 126-186 in W. W. Hay, ed. Studies in Paleo-Ocean- ography. Society of Economic Paleontology and Mineralogy, Special Publication, no. 20. BERMINGHAM, E. B., AND H. A. LESSIOS. 1993. Rate variation of protein and mtDNA evolution as revealed by sea urchins sep- arated by the Isthmus of Panama. Proc. Natl. Acad. Sci. USA 90:2734-2738. BERNASCONI, I. 1955. Equinoideos y asteroideos de la colecci6n del Instituto Oceanographico de la Universidad de San Pablo. Bol. Inst. Oceanogr. S. Paulo 6:51-77. BIERNBAUM, C. K. 1996. Biogeography of coastal and anchialine amphipods of Ascension Island, south Atlantic Ocean. J. Nat. Hist. 30:1597-1615. BRIGGS, J. C. 1974. Marine zoogeography. McGraw-Hill, New York. CHARLESWORTH, B. 1998. Measures of divergence between pop- ulations and the effect of forces that reduce variability. Mol. Biol. Evol. 15:538-543. CHAVEZ, F. P., AND R. C. BRUSCA. 1991. The Galapagos Islands and their relation to oceanographic processes in the tropical Pacific. Pp. 9-33 in M. J. James, ed. Galapagos marine inver- tebrates. Plenum Press, New York. CHESHER, R. H. 1966. The R/V PILLSBURY deep-sea biological expedition to the Gulf of Guinea, 1964-65. Vol. 10. Report on the Echinoidea collected by the R/V PILLSBURY in the Gulf of Guinea. Stud. Trop. Oceanogr. 4:209-223. CHRISTIE, D. M., R. A. DUNCAN, A. R. McBIRNEY, M. A. RICHARDS, W. M. WHITE, K. S. HARPP, AND C. G. Fox. 1992. Drowned islands downstream from the Galapagos hotspot imply extended speciation times. Nature 355:246-248. CLARK, A. M., AND J. COURTMAN-STOCK. 1976. The echinoderms of southern Africa. British Museum (Natural History), London. CLARK, H. L. 1925. A catalogue of the recent sea urchins (Echi- noidea) in the collection of the British Museum (Natural His- tory). Oxford Univ. Press, London. COATES, A. G., AND J. A. OBANDO. 1996. The geologic evolution of the Central American Isthmus. Pp. 21-56 in J. B. C. Jackson, A. G. Coates, and A. Budd, eds. Evolution and environment in tropical America. Univ. of Chicago Press, Chicago. This content downloaded on Thu, 17 Jan 2013 14:37:51 PM All use subject to JSTOR Terms and Conditions 816 H. A. LESSIOS ET AL. CRONIN, T. M., AND H. J. DOWSETT. 1996. Biotic and oceanographic response to the Pliocene closing of the Central American Isth- mus. Pp. 76-104 in J. B. C. Jackson, A. F Budd, and A. G. Coates, eds. Evolution and environment in tropical America. Univ. of Chicago Press, Chicago. CUTRESS, B. M. 1980. Cretaceous and Tertiary Cidaroida (Echi- nodermata: Echinoidea) of the Caribbean area. Bull. Am. Pa- leontol. 77:1-221. DIESTER-HAASS, L., AND H. J. SCHRADER. 1979. Neogene coastal upwelling history off northwest and southwest Africa. Mar. Geol. 29:39-53. DODERLEIN, L. 1887. Die Japanischen Seeigeln. I. Familie Cidar- idae und Saleniidae. Vol. 1. E. Schweizerbart'sche Verlagshan- dlung, Stuttgart. DURHAM, J. W. 1966. Evolution among the Echinoidea. Biol. Rev. 41:368-391. EDWARDS, A. 1990. Fish and fisheries of Saint Helena Island. Cen- tre for Tropical Coastal Management Studies, Univ. of Newcastle upon Tyne, Newcastle upon Tyne, U.K. EDWARDS, A., AND R. LUBBOCK. 1983. The ecology of Saint Paul's Rocks (equatorial Atlantic). J. Zool. Lond. 200:51-69. EKMAN, S. 1953. Zoogeography of the sea. Sidgwick and Jackson Ltd., London. EMLET, R. B. 1988. Larval form and metamorphosis of a "prim- itive" sea urchin, Eucidaris thouar-si (Echinodermata: Echinoi- dea: Cidaroida), with implications for developmental and phy- logenetic studies. Biol. Bull. 174:4-19. FELL, H. B. 1966. Cidaroids. Pp. 312-366 In R. C. Moore, ed. Treatise on invertebrate paleontology. Geological Society of America and Univ. of Kansas Press, Lawrence. GARTH, J. S. 1991. Taxonomy, distribution, and ecology of Gal- apagos Brachyura. Pp. 123-145 in M. J. James, ed. Galapagos marine invertebrates. Plenum Press, New York. GLYNN, P. W. 1994. State of coral reefs in the Galapagos Islands: natural vs anthropogenic mpacts. Mar. Pollut. Bull. 29:131-140. GLYNN, P. W., AND G. M. WELLINGTON. 1983. Corals and coral reefs of the Galapagos Islands. Univ. of California Press, Berke- ley. GLYNN, P. W., G. M. WELLINGTON, AND C. BIRKELAND. 1979. Coral reef growth in the Galapagos: limitation by sea urchins. Science 203:47-48. GRIGG, R. W., AND R. HEY. 1992. Paleoceanography of the tropical eastern Pacific Ocean. Science 255:172-178. HALLAM, A. 1994. An outline of Phanerozoic biogeography. Ox- ford Univ. Press, Oxford. HASEGAWA, M., H. KISHINO, AND T. YANO. 1985. Dating of the human-ape splitting by a molecular clock of mitochondrial DNA. J. Mol. Evol. 22:160-174. HAUG, G. H., AND R. TIEDEMANN. 1998. Effect of the formation of the Isthmus of Panama on Atlantic Ocean thermohaline cir- culation. Nature 393:673-678. HERTLEIN, L. G. 1963. Contribution to the biogeography of Cocos Island, including a bibliography. Proc. Calif. Acad. Sci. 32:219- 289. HUDSON, R. R., M. SLATKIN, AND W. P. MADDISON. 1992. Esti- mation of levels of gene flow from DNA sequence data. Genetics 132:583-589. JACOBS, H. T., D. J. ELLIOT, B. M. VEERABHADRACHARYA, AND A. FARQUHARSON. 1988. Nucleotide sequence and gene organi- zation of sea urchin mitochondrial DNA. J. Mol. Biol. 202:185- 217. JAMES, M. J. 1984. A new look at evolution in the Galapagos: evidence from the late Cenozoic marine molluscan fauna. Biol. J. Linn. Soc. 21:77-95. KAY, E. A. 1991. The marine mollusks of the Galapagos: deter- minants of insular marine faunas. Pp. 235-252 in M. J. James, ed. Galapagos marine invertebrates. Plenum Press, New York. KELLER, G., C. E. ZENKER, AND S. M. STONE. 1989. Late Noegene history of the Pacific-Caribbean gateway. J. So. Am. Earth Sci. 2:73-108. KIMURA, M. 1980. A simple method for estimating evolutionary rates of base substitutions through comparative studies of nu- cleotide sequences. J. Mol. Evol. 16:111-120. LESSIOS, H. A. 1979. Use of Panamanian sea urchins to test the molecular clock. Nature 280:599-601. 1981. Divergence in allopatry: molecular and morpholog- ical differentiation between sea urchins eparated by the Isthmus of Panama. Evolution 35:618-634. . 1990. A program for calculating Nei's genetic distances and their jackknifed confidence intervals. J. Hered. 81:490. . 1998. The first stage of speciation as seen in organisms separated by the Isthmus of Panama. Pp. 186-201. in D. Howard and S. Berlocher, eds. Endless forms: species and speciation. Oxford Univ. Press, Oxford. LESSIOS, H. A., B. D. KESSING, G. M. WELLINGTON, AND A. GRAY- BEAL. 1996. Indo-Pacific echinoids in the tropical eastern pa- cific. Coral Reefs 15:133-142. LESSIOS, H. A., B. D. KESSING, AND D. R. ROBERTSON. 1998. Mas- sive gene flow across the world's most potent marine biogeo- graphic barrier. Proc. R. Soc. Lond. Ser. B Biol. Sci. 265:583- 588. Li, W.-H. 1993. Unbiased estimation of the rates of synonymous and nonsynonymous substitution. J. Mol. Evol. 36:96-99. LONGHURST, A. R. 1962. A review of the oceanography of the Gulf of Guinea. Bull. Inst. Franc. d'Afrique Noire. 24:633-663. LUBBOCK, R. 1980. The shore fishes of Ascension Island. J. Fish. Biol. 17:283-303. LUYENDYK, B. P., D. FORSYTH, AND J. D. PHILLIPS. 1972. Exper- imental approach to the paleocirculation of the oceanic surface waters. Geol. Soc. Am. Bull. 83:2649-2664. MAIER-REIMER, E., U. MIKOLAJEWICZ, AND T. CROWLEY. 1990. Ocean general circulation model sensitivity experiment with an open central American isthmus. Paleoceanography 5:349-366. MALUF, L. Y. 1988. Composition and distribution of the central eastern Pacific echinoderms. Nat. Hist. Mus. Los Angeles Co. Tech. Rept. 2:1-242. . 1991. Echinoderm fauna of the Galapagos Islands. Pp. 345-367 in M. J. James, ed. Galapagos marine invertebrates. Plenum Press, New York. MANNING, R. B., AND F. A. CHACE. 1990. Decapod and stomatopod Crustacea from Ascension Island, South Atlantic Ocean. Smiths. Contrib. Zool. 503:1-91. MAYR, E. 1954. Geographic speciation in tropical echinoids. Evo- lution 8:1-18. MCKENZIE, K. G. 1967. The distribution of Caenozoic marine os- tracoda from the Gulf of Mexico to Australasia. Syst. Assoc. Publ. 7:219-238. MCPHERSON, B. F 1968. Contributions to the biology of the sea urchin Eucidaris tribuloides (Lamarck). Bull. Mar. Sci. 18:400- 443. METCALF, W. G., A. D. VOORHIS, AND M. G. STALEUP. 1962. The Atlantic Equatorial Undercurrent. J. Geophys. Res. 67:2499- 2508. MITCHELL-THoMt, R. C. 1982. The geological settings and char- acteristics of the Atlantic islands. Acta Geol. Acad. Sci. Hun- garicae 25:395-420. MORTENSEN, T. 1928-1951. A monograph of the Echinoidea. C. A. Reitzel, Copenhagen. . 1932. New contributions to the knowledge of the cidarids I-II. Danske Videnskabernes Selskabs Skrifter, Naturvidenska- beling og mathematisk Afdeling 4:1-39. . 1933. The echinoderms of St. Helena (other than crinoids). Videnskabelige Meddelesler fra Dansk naturhistorisk Forening i Kjobenhavn 93:401-472. 1936. Echinoidea and Ophiuroidea. "Discovery" Reports 12:199-348. MUELLER, L. D., AND F. J. AYALA. 1982. Estimation and interpre- tation of genetic distance in empirical studies. Genet. Res. Camb. 40:127-137. NEI, M. 1978. Estimation of average heterozygosity and genetic distance from a small number of individuals. Genetics 89:583- 590. PALUMBI, S. R. 1996. What can molecular genetics contribute to marine biogeography? an urchin's tale. J. Exper. Mar. Biol. Ecol. 203:75-92. . 1997. Molecular biogeography of the Pacific. Pp. 91-96 This content downloaded on Thu, 17 Jan 2013 14:37:51 PM All use subject to JSTOR Terms and Conditions PHYLOGEOGRAPHY OF EUCIDARIS 817 in H. A. Lessios and I. G. Macintyre, eds. Proceedings of the 8th International Coral Reef Symposium, Smithsonian Tropical Research Institute, Balboa, Panama. PALUMBI, S. R., G. GRABOWSKY, T. DUDA, L. GEYER, AND N. TACH- INO. 1997. Speciation and population genetic structure in trop- ical Pacific sea urchins. Evolution 51:1506-1517. PAMILO, P., AND 0. BIANCHI. 1993. Evolution of the Zfx and Zfy genes: rates and interdependence between the genes. Mol. Biol. Evol. 10:271-281. PAWSON, D. L. 1978. The echinoderm fauna of Ascension Island, south Atlantic Ocean. Smiths. Contr. Mar. Sci. 2:1-31. PICCOLI, G., S. SARTORI, AND A. FRACHINO. 1987. Benthic molluscs of shallow Tethys and their destiny. Pp. 333-373 In K. G. McKenzie, ed. Shallow Tethys. A. A. Balkema, Rotterdam. POR, F D. 1989. The legacy of Tethys: an aquatic biogeography of the Levant. Kluwer, Dordrecht. RASSMAN, K. 1997. Evolutionary age of the Galapagos iguanas predates the age of the present Galapagos Islands. Mol. Phylog. Evol. 7:158-172. ROBBA, E. 1987. The final occlusion of Tethys: its bearing on Mediterranean benthic molluscs. Pp. 405-426 in K. G. Mc- Kenzie, ed. Shallow Tethys. A. A. Balkema, Rotterdam. ROBERTSON, D. R., AND G. R. ALLEN. 1996. Zoogeography of the shorefish fauna of Clipperton Atoll. Coral Reefs 15:121-131. ROGL, F., AND F F STEININGER. 1984. Neogene Paratethys, Med- iterranean and Indo-Pacific seaways. Implications for the paleo- biogeography of marine and terrestrial biotas. Pp. 171-200 in P. Brenchley, ed. Fossils and climate. Wiley, New York. ROSEN, B. R. 1984. Reef coral biogeography and climate through the Late Cainozoic: just islands in the sun or a critical pattern of islands? Pp. 201-262 in P. Brenchley, ed. Fossils and climate. Wiley, New York. ROSEN, B. R., AND A. B. SMITH. 1988. Tectonics from fossils? analysis of reef-coral and sea-urchin distributions from late Cre- taceous to Recent, using a new method. Pp. 275-306 in M. G. Audley-Charles and A. Hallam, eds. Gondwana and Tethys. Ox- ford Univ. Press, Oxford. ROSEWATER, J. 1975. An annotated list of the marine mollusks of Ascension Island, south Atlantic Ocean. Smiths. Contrib. Zool. 189: 1-41. RUGGIERI, G. 1967. The Miocene and later evolution of the Med- iterranean Sea. Syst. Assoc. Publ. no. 7:283-290. SAITOU, N., AND M. NEI. 1987. The neighbor-joining method: a new method for reconstructing phylogenetic trees. Mol. Biol. Evol. 4:406-425. SCHELTEMA, R. S. 1995. The relevance of passive dispersal for the biogeography of Caribbean mollusks. Am. Malacol. Bull 11:99- 115. SEUTIN, G., B. N. WHITE, AND P. T. BOAG. 1991. Preservation of avian blood and tissue samples for DNA analyses. Can. J. Zool. 69:82-90. SHANNON, L. V. 1985. The Benguela ecosystem. Part I. Evolution of the Benguela physical features and processes. Oceanogr. Mar. Bio. Annu. Rev. 23:105-182. SIESSER, W. G. 1980. Late Miocene origin of the Benguela up- welling system off northern Namibia. Science 208:283-285. SMITH, A. 1984. Echinoid paleobiology. Allen and Unwin, London. SMITH, A. B., AND C. W. WRIGHT. 1989. British Cretaceous echi- noids. Part 1. General introduction and Cidaroidea. Monogr. Pa- laeontogr. Soc. Publ. 578:1-101. STRIMMER, K., AND A. VON HAESELER. 1996. Quartet Puzzling: a quartet maximum-likelihood method for reconstructing tree to- pologies. Mol. Biol. Evol. 13:964-969. TSUCHIYA, M. 1974. Variation of the surface geostrophic flow in the eastern intertropical Pacific Ocean. Fish. Bull. US 72:1075- 1086. VERMEIJ, G. J. 1978. Biogeography and adaptation. Harvard Univ. Press, Cambridge, MA. . 1987. The dispersal barrier in the tropical Pacific: impli- cation for molluscan speciation and extinction. Evolution 41: 1046-1058. WALKER, B. W. 1966. The origins and affinities of the Galapagos shorefishes. Pp. 172-174 in R. I. Bowman, ed. The Galapagos. Univ. of California Press, Los Angeles. WELLINGTON, G. M. 1997. Field guide to the corals and coral reefs of the Galapagos Islands, Ecuador. Pp. 185-202. in H. A. Lessios and I. G. Macintyre, eds. Proceedings of the 8th International Coral Reef Symposium, Smithsonian Tropical Research Insti- tute, Balboa, Panama. WINTERBOTTOM, R., AND D. A. McLENNAN. 1993. Cladogram ver- satility: evolution and biogeography of acanthuroid fishes. Evo- lution 47:1557-1571. WYLES, J. S., AND V. M. SARICH. 1983. Are the Galapagos iguanas older than the Galapagos? Molecular evolution and colonization models for the archipelago. Pp. 177-185 in R. I. Bowman, M. Berson, and A. E. Levinton, eds. Patterns of evolution in Gal- apagos organisms. American Association for the Advancement of Science, Pacific Division, San Francisco, CA. WYRTKI, K. 1965. Surface currents of the eastern equatorial Pacific Ocean. Inter-Amer. Trop. Tuna Comm. Bull. 9:270-304. 1966. Oceanography of the eastern equatorial Pacific Ocean. Oceanogr. Mar. Bio. Annu. Rev. 4:33-68. . 1967. Circulation and water masses in the eastern equa- torial Pacific Ocean. Int. J. Oceanol. Limnol. 1:117-147. Corresponding Editor: S. Palumbi This content downloaded on Thu, 17 Jan 2013 14:37:51 PM All use subject to JSTOR Terms and Conditions