TREE-1666; No.

of Pages 8

Review

The origins of tropical marine biodiversity

Brian W. Bowen1, Luiz A. Rocha2, Robert J. Toonen1, Stephen A. Karl1, and the ToBo Laboratory*
1

Hawaii Institute of Marine Biology, School of Ocean and Earth Science and Technology, University of Hawaii, P.O. Box 1346, Kaneohe, HI 96744, USA 2 California Academy of Sciences, 55 Music Concourse Drive, San Francisco, CA 94118, USA

Recent phylogeographic studies have overturned three paradigms for the origins of marine biodiversity. (i) Physical (allopatric) isolation is not the sole avenue for marine speciation: many species diverge along ecological boundaries. (ii) Peripheral habitats such as oceanic archipelagos are not evolutionary graveyards: these regions can export biodiversity. (iii) Speciation in marine and terrestrial ecosystems follow similar processes but are not the same: opportunities for allopatric isolation are fewer in the oceans, leaving greater opportunity for speciation along ecological boundaries. Biodiversity hotspots such as the Caribbean Sea and the Indo-Pacic Coral Triangle produce and export species, but can also accumulate biodiversity produced in peripheral habitats. Both hotspots and peripheral ecosystems benet from this exchange in a process dubbed biodiversity feedback. The enigma of speciation in the sea Over half a century ago Ernst Mayr posed the question of whether speciation is fundamentally different in the sea [1]. On the basis of studies of sea urchins, he concluded that the process was the same in marine and terrestrial faunas, but this controversy ared across ensuing decades. Much of the debate centered on the relative importance of genetic isolation versus dispersal in a scientic era in which physical (allopatric) barriers were regarded as the primary, if not exclusive, starting point for speciation [2,3]. There are few obvious physical barriers in the ocean, and dispersal is extensive in this transglobal aquatic medium; two factors that should reduce opportunities for allopatric speciation [4]. Yet tropical reefs host biodiversity that rivals that in rainforests. Coral reefs cover less than 0.1% of the ocean oor [5], but their sh communities include approximately one-third of the recognized marine
Corresponding author: Bowen, B.W. (bbowen@hawaii.edu). * Members of the ToBo (Toonen/Bowen) Laboratory who contributed to this paper are Matthew T. Craig (Department of Marine Science and Environmental Studies, University of San Diego, San Diego, CA 92110, USA), Joseph D. DiBattista (Red Sea Research Center, King Abdullah University of Science and Technology, Thuwal, Saudi Arabia), Jeff A. Eble (Center for Environmental Diagnostics and Bioremediation, University of West Florida, Pensacola, FL 32514 USA), Michelle R. Gaither2, Derek Skillings1, and Chris J. Bird (Department of Life Sciences, Texas A&M University Corpus Christi, 6300 Ocean Drive, Corpus Christi, TX, 78412 USA). Keywords: biodiversity feedback; biogeography; center of accumulation; center of speciation; ecological opportunity; marine speciation; phylogeography. 0169-5347/$ see front matter 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.tree.2013.01.018

species. In these circumstances it is impossible to reconcile the prevailing beliefs about speciation by allopatric isolation with the plethora of closely related species living together in the sea. If cessation of gene ow by geographic isolation is the key to terrestrial speciation, then intuitively the processes leading to speciation should be different between marine and terrestrial habitats. Most marine organisms move little as juveniles and adults, so connectivity must be through pelagic (oceanic) stages such as larvae, and in some cases this dispersal can be extensive. Whereas some reef fauna show self-recruitment over small geographic ranges [6], the majority have oceanic dispersing larvae that can contribute substantially to connectivity among regions [7]. Marine biotas show a broad range of dispersal capabilities [8], from crawl-away larvae to pelagic dispersal that is at least an order of magnitude greater than the Glossary
Biodiversity feedback: biogeographic model with biodiversity hotspots acting as both Centers of Speciation and Centers of Accumulation and/or Overlap. Biodiversity hotspot: region with especially high levels of biodiversity as measured by species numbers and levels of endemism. Conservation of biodiversity hotspots should aim at preserving not just species diversity but also the mechanisms underlying biodiversity production. Center of Accumulation model: this model suggests that the high number of species in the Coral Triangle results from speciation in peripheral locations with subsequent dispersal of novel taxa into the Coral Triangle. The long history of the Pacific archipelagos, some of which date to the Cretaceous, the possibility of isolation in these peripheral habitats, and current and wind patterns that favor dispersal towards the Coral Triangle have been suggested as a mechanism. Center of Overlap model: this model suggests that the high species diversity in the Coral Triangle is in part due to overlap of distinct faunas from the Pacific and Indian Oceans. The hypothesis is based on the premise that the isolating mechanism is the Indo-Pacific Barrier, which separates the Pacific and Indian Oceans during low sea-level stands, with the faunas of these oceans diverging during periods of isolation. Center of Speciation model: also known as the Center of Origin, this biogeographic model suggests that tropical biodiversity hotspots including the Coral Triangle are exporters of species, possibly driven by the fracture of populations that result from geologic complexity and habitat heterogeneity coupled with intense competition. Coral Triangle: epicenter of marine biodiversity, with over 2700 species of shore fishes and 600 species of corals, that extends from the Philippines, eastern Indonesia and New Guiana southeast to the Solomon Islands (see Figure 1 in main text). Evolutionary graveyard: the premise that depauperate peripheral habitats act solely as sinks of biodiversity and contribute little to overall species richness. Gametic recognition proteins (GRP): cell surface proteins that mediate the interactions between gamete cells with incompatibility, resulting in unsuccessful fertilization. Rapid evolution of GRPs is thought to be an important driver of speciation in some taxa.

Trends in Ecology & Evolution xx (2012) 18

TREE-1666; No. of Pages 8

Review
capabilities of pollen-dispersing terrestrial biotas [9]. For example, moray eels (family Muraenidae) show high population genetic connectivity across two-thirds of the planet, from Africa to Central America [10], unicornshes (Naso spp.) show no genetic isolation across the Indian and Pacic Oceans [11], and even the diminutive pygmy angelshes (genus Centropyge) show high genetic connectivity across thousands of kilometers [12]. To the extent that dispersal and gene ow inuence speciation, the process cannot be the same on land and sea, or must work on vastly different scales. Here we review recent work indicating that (i) speciation in the sea can proceed without absolute barriers to gene ow, (ii) peripheral habitats such as the Hawaiian Archipelago and Red Sea can produce and export new species, and (iii) exchange of biota among extremely competitive hotspots and depauperate peripheral areas is a process that enhances species numbers and ecosystem resilience in both areas, which we term the biodiversity feedback model (see Glossary) [13]. This synthesis arises primarily from efforts to resolve the origins of marine biodiversity hotspots, including three apparently overlapping hypotheses reviewed in the next section. Biodiversity hotspots John Briggs has been the primary advocate of marine centers of speciation, among which the Coral Triangle in the Indo-Pacic (also known as Indo-Australian Archipelago or Indo-Malayan Triangle, the area between Indonesia, New Guinea, and the Philippines in Figure 1) and the Caribbean Sea in the Atlantic are cradles for new species that radiate out and replace older species [14]. Under this model, intense competition in highly diverse habitats yields new species with greater tness than their predecessors. A combination of physical and ecological partitioning is probably the most effective means to establish new species. Therefore, isolated populations within the Coral Triangle might be the precursors of speciation and subsequent radiations [1517]. The distribution of reef shes is consistent with this model [18] and it is apparent that much of the biodiversity in peripheral archipelagos has arrived from the central

Trends in Ecology & Evolution xxx xxxx, Vol. xxx, No. x

Indo-Pacic [19]. Recent evidence indicates that two reefassociated sharks (Sphyrna lewini and Triaenodon obesus) radiated from the central Indo-Pacic to attain broad or global distributions [20,21]. Two new reviews provide concordant assessments for the Coral Triangle hotspot, including a long history of species accumulation beginning in the Miocene [2012 million years ago (mya)] and subsequent export of biodiversity in the Pliocene, Pleistocene, and Holocene (7 mya to the present) [22,23]. A corollary of the Center of Speciation hypothesis is that areas peripheral to biodiversity hotspots are evolutionary graveyards, where colonizers arrive and evolve into endemic species, but never radiate beyond their new home. This is supported by studies of the isolated Hawaiian fauna, including sh, corals, and other invertebrates [2426]. Recent studies have prompted a reappraisal of this corollary, which we examine below. A primary alternative to the Center of Speciation hypothesis is a Center of Accumulation, in which new species arise in peripheral habitats such as oceanic archipelagos and accumulate in the center of the range because of prevailing currents [27,28]. Support for this model comes from several phylogeographic studies (see below). Budd and Pandol note that morphological novelty is highest at the edge of coral species distributions, consistent with the Center of Accumulation model [29]. A related model is the Center of Overlap, in which isolated faunas of the Indian and Pacic Oceans are codistributed in the area of overlap, including the Coral Triangle. The biogeographic case for Center of Overlap includes the many sister species with Pacic and Indian Ocean distributions that overlap at the boundary of these oceans [30]. Gaither et al. observed two mtDNA lineages in the grouper Cephalopholis argus corresponding to Indian and Pacic Oceans but overlapping in the boundary regions between these oceans (including the Coral Triangle; Figure 1) [31]. Overall, there is substantial evidence in support of all three processes according to recent phylogeographic studies of tropical marine fauna [13,32,33]. Whereas the Center of Speciation model involves speciation without complete

30

30 30 60 120 180 120

TRENDS in Ecology & Evolution

Figure 1. Biogeographic provinces of the tropical West Atlantic and Indo-Pacific (indicated in blue) [23,93]. The Biodiversity Feedback model indicates that biodiversity flows out from the Coral Triangle (light blue in central Indo-Pacific) to peripheral habitats in Hawaii, the East Pacific, Red Sea, and elsewhere. In the Atlantic, Caribbean fauna can colonize the depauperate coastline of Brazil. Species diverging in these peripheral habitats can subsequently radiate back out to enhance overall biodiversity [22,23].

TREE-1666; No. of Pages 8

Review
geographic isolation (allopatry), the other two models include some level of allopatric isolation, which raises the question as to whether isolation plays a leading role in marine speciation. Speciation without geographic barriers Speciation via allopatry, as proposed by Dobzhansky [34] and Mayr [2], continues to be the dominant paradigm in evolutionary biology [35]. By contrast, molecular phylogenies of marine organisms provide many examples of sister taxa with distributions that are adjacent (parapatric), overlapping, or even identical (sympatric). In a growing body of literature, non-allopatric speciation is inferred from phylogenetic data combined with distribution data and ecological partitions (Box 1). Examples of

Trends in Ecology & Evolution xxx xxxx, Vol. xxx, No. x

sympatric sister species have been documented in a diversity of taxa including East Pacic shes [36], North Atlantic snails [37], and Caribbean sponges [38]. Among the most intriguing examples are sympatric species that seem to maintain separate identities despite ongoing gene ow, including Pacic pygmy angelshes (Centropyge avissima complex) [12], Caribbean hamlets (genus Hypoplectrus) [39], and sea stars (Linckia multiora and Linckia laevigata) [40] (Figure 2). These may represent the earliest stages of speciation. In addition to overlapping sister species, numerous examples have emerged of sister species with adjacent distributions. Caribbean gobies (genus Elacatinus) have several color morphs with parapatric distributions, for which coloration differences might facilitate assortative

Box 1. Cases of non-allopatric speciation

Caribbean and East Pacific reef fish, collectively known as grunts (genus Haemulon), show numerous examples of sympatric sister species [50]. Vocalization may be key to mate recognition in the darkness as these fishes spawn at night. It is suspected that sexual selection on the basis of vocal cues promotes reproductive isolation among sister species. Photo: L.A. Rocha. The brooding coral Seriatopora histrix shows genetic partitions between adjacent habitat types on the Great Barrier Reef [98], prompting the authors to conclude that adaptation to the local environment has caused ecological divergence of distinct genetic groups within S. hystrix. Photo: J. Sprung. West Pacific reef fishes include several examples of sympatric sister species characterized by habitat shifts, including gobies (Gobiodon spp.) [99]. In this case, sympatric speciation has been accompanied by a host shift between Acropora corals. Photo: P.L. Munday. Hawaiian limpets (Cellana spp.) comprise three sympatric sister species that partition along ecological lines into upper, middle, and lower intertidal zones. Habitat partitioning is thought to be caused by desiccation, thermal tolerance, and/or predation avoidance mechanisms, and reproductive isolation may be accomplished by the timing of gametic release [91]. Photo: C.L. Bird. Nudibranchs (genus Phestilla) form a close bond with their scleractinian coral host, feeding, reproducing, and requiring host-specific chemical cues for settlement. Phylogenetic analyses revealed several incidences of host-shifting accompanied by speciation [89]. Photo: R. Williams.

Sister species of brown algae (genus Fucus) segregate along a steep intertidal gradient despite ongoing hybridization. Common garden experiments show that physiological resistance to desiccation in each species is commensurate with intertidal exposure time [94]. Photo: K.R. Nicastro. Caribbean wrasses (Halichoeres spp.) reveal intraspecific genetic partitions between adjacent and ecologically distinct habitats, but high genetic connectivity between similar habitats separated by thousands of kilometers [95]. Photo: L.A. Rocha. The Caribbean sponge Chondrilla caribensis has two ecotypes, one proposed ancestral ecotype adapted to reef habitats and another specialized for mangrove habitats. These ecotypes, long described as a single species, have been distinguished on the basis of morphology and genetics [96]. Photo: K. Ruetzler. The Caribbean basslet Gramma dejongi has a restricted geographic range completely encompassed by the widespread sister species G. loreto. In describing this species, Victor and Randall note that G. dejongi has a narrower depth range than G. loreto [97]. Photo: B. Victor.

TREE-1666; No. of Pages 8

Review

Trends in Ecology & Evolution xxx xxxx, Vol. xxx, No. x

(a)

(b)

(c)

(d)

(e)

(f)

TRENDS in Ecology & Evolution

Figure 2. Sister species with overlapping distributions that maintain striking color differences despite apparent gene flow: (a) Hypoplectrus puella and (b) Hypoplectrus gemma in the Caribbean Sea [39]; (c) Centropyge vrolikii and (d) Centropyge flavissima in the West Pacific [12]; and (e) Linckia multiflora and (f) Linckia laevigata in the West Pacific [40]. Photos: L.A. Rocha and G. Paulay.

mating at range boundaries [41]. The Dascyllus trimaculatus species complex of four closely related Pacic damselshes has parapatric distributions [42]. Particularly interesting cases of parapatry include highly mobile marine vertebrates with ecotypes that segregate into adjacent (often inshore and offshore) morphotypes, a phenomenon observed in marine mammals [43,44], sharks [45], and manta rays [46]. These species are all capable of vast dispersal and yet retain genetic partitions along ecological boundaries. These and many other examples have prompted researchers to resurrect ecological factors as primary drivers in marine speciation [47]. This Darwinian view of selection driving speciation is a marked departure from the era of vicariant biogeography, when speciation was dened in terms of geography and physical isolation [3]. Clearly, a single model of speciation via allopatric isolation
4

is insufcient to describe the diversity of species in the sea. The combination of hard, soft, or intermittent barriers (such as oceanic currents, water masses, expanses of deep ocean, or land masses exposed during glaciation) and ecological differences between adjacent areas represents a middle ground between speciation in sympatric or allopatric isolation, a very fertile circumstance for speciation in the sea [4,48]. A counterargument to these cases of apparent sympatric speciation is based on allopatric speciation with subsequent range expansion and secondary contact. Under this scenario, speciation occurs in isolation but then species ranges expand to overlap [49]. This explanation of historical changes in geographic distribution is very difcult to disprove, and is almost certainly correct in some cases. Isolation followed by secondary contact can explain diversication within the Coral Triangle for less dispersive

TREE-1666; No. of Pages 8

Review
species (e.g., the clownsh Amphiprion ocellaris) [16]. It is harder, however, to t this hypothesis to recently derived species (such as the grunts, Haemulon spp.) in which many sister species have entirely overlapping distributions [50]. Several research groups have concluded that sympatric (or parapatric) speciation is faster than the allopatric alternative [35,51], averaging approximately 200 000 years versus 2.7 million years in Drosophila, presumably because selection can establish a reproductive barrier more rapidly than random genetic drift due to physical isolation [52,53]. Rapid evolution of gametic recognition proteins (bindins) can facilitate the nal stages of reproductive isolation [54]. Examples of rapid speciation include marine snails [55], seastars [56], and marine shes [57]. Recent modeling studies indicate that the more rapid speciation in sympatry or parapatry is likely driven by a combination of sexual selection (mate recognition and assortative mating) combined with adaptive natural selection for alternative habitats [58]. Peripheral habitats can export new species The Center of Speciation versus Center of Accumulation and/or Overlap hypotheses cast oceanic islands in opposing roles. In the Center of Speciation model, islands are evolutionary dead ends, whereas in the Center of Accumulation and/or Overlap, islands are engines of biodiversity production. Recent studies have evaluated these alternatives. Although some ocean island endemics appear to be relic populations of previously widespread species [19], many have become established recently (neo-endemic; e.g., Thalassoma spp.) [32], with their presence demonstrating evolutionary innovation around oceanic islands. Accumulated evidence indicates that areas peripheral to biodiversity hotspots are exporting genetic and biological diversity [59]. The ember parrotsh (Scarus rubroviolaceus) occurs throughout the Indo-Pacic, but population genetic data indicate that larval export from Hawaii is an order of magnitude greater than import [60], and this ratio is even higher for the deepwater crimson jobsh Pristipomoides lamentosus [61] and the lollysh sea cucumber Holothuria atra [62]. The bullethead parrotsh (Chlorurus sordidus) has three distinct mtDNA lineages corresponding to the Indian Ocean, the West Pacic, and Hawaii. However, sh collected in the Northern Marianas Islands (westward and down-current from Hawaii) have the Hawaiian lineage, indicating a period of extended Hawaiian isolation followed by more recent larval export [63]. Migration estimates for the IndoPacic yellow tang (Zebrasoma avescens) are similar to those for the ember parrotsh, with a prevailing pattern of larval export from Hawaii [64]. Moreover, yellow tang ecology, historical demography, and phylogenetics point to a Hawaiian origin for this widely distributed reef sh. Hence, Hawaii is both a source and recipient of IndoPacic marine diversity. The Red Sea can also export unique genetic lineages to adjacent regions. Despite a volatile geological history characterized by intermittent isolation and multiple salinity crises, mtDNA distributions of widespread reef sh species indicate that some species originate in the Red Sea with subsequent export to the Indian Ocean [65].

Trends in Ecology & Evolution xxx xxxx, Vol. xxx, No. x

Isolated archipelagos and peripheral seas are not the only sources conveying species into areas of higher biodiversity. The continental coastlines of the Indo-Pacic and Atlantic can also provide evolutionary novelties. The parrotsh genus Scarus includes two basal lineages (Scarus zufar and Scarus collana) endemic to the coastlines of the northern Indian Ocean and Red Sea [66]. The continental reefs of the tropical eastern Pacic have contributed at least 23 species to the Central Pacic [67]. Much of the reef fauna of the eastern tropical Pacic are derived from the central Indo-Pacic, involving migration events across the vast open ocean between the Central and East Pacic (Figure 1). However, phylogeographic studies indicate recent or ongoing connections in both directions [68]. In the tropical Atlantic, the Caribbean is recognized as a biodiversity hotspot with approximately 814 reef shes, and the Brazilian coastline is a depauperate outpost of this fauna with approximately 471 species [69]. Most of the Brazilian fauna is derived from the Caribbean, but Brazil also exports both genetic diversity and sh species back into the Caribbean [13]. These studies demonstrate that isolated islands, marginal seas, and peripheral coastlines are not evolutionary graveyards for marine fauna. We anticipate that with more investigations, the number of examples of biodiversity export from peripheral areas will grow. Clearly, biodiversity production in tropical seas does not fall easily into a model of a Center of Speciation, Accumulation, or Overlap, but requires a more inclusive model that incorporates all of the observed patterns. Biodiversity feedback: a new paradigm In recent years a number of phylogeographic studies have tested whether the youngest lineages or species are in biodiversity hotspots (favoring Center of Speciation) or in peripheral habitats (favoring Center of Accumulation and/or Overlap). The results have been mixed. Barber et al. support a Center of Speciation on the basis of the genetic architecture of mantis shrimp (Stomatopods) [15], and similar patterns have been reported for mangroves (Rhizophoraceae) [70] and turban snails (Gastropoda) [71]. Teske et al. support the Center of Accumulation model on the basis of seahorse (Hippocampus spp.) phylogenetics [72]. In a survey of eight shes, four corals, and one mollusk, Halas and Winterbottom observed evidence for both hypotheses [73]. Phylogeographic patterns in wrasses (Halichoeres spp. and Thalassoma spp.) also support both models [32,74]. These studies laid the foundations for the current synthesis, with the recognition that these models are not mutually exclusive. The new biodiversity feedback model [13] incorporates elements of all three previous models (Box 2). There is no question that allopatric speciation occurs in the sea, and many examples have emerged from phylogeographic studies [75]. It is also clear that rare dispersal events to oceanic archipelagos, followed by isolation and speciation, are an important evolutionary process responsible for the high level of endemism in remote archipelagos [76,77]. The paradigm shift here is that these marine species can sometimes radiate further and escape the evolutionary dead end of peripheral isolation. In Hawaii,
5

TREE-1666; No. of Pages 8

Review

Trends in Ecology & Evolution xxx xxxx, Vol. xxx, No. x

Box 2. The false dichotomy between competing hypotheses in ecology and evolution
The scientific method mandates testing of alternate hypotheses, with predictions of each hypothesis evaluated using empirical data. In many fields of science, testing of alternative hypotheses is a robust approach. For instance, the physical particles that comprise atoms have a positive, negative, or neutral charge, and none of these labels can apply simultaneously. In biology, a more nuanced approach is required, because multiple explanations are needed to explain complex biological patterns. For example, hypotheses for the navigation of migratory birds involve celestial clues (stars and sun) or sensing of the geomagnetic field of the earth, when in fact each is correct for individual species and some species use both [100]. Likewise, the Center of Speciation, Center of Accumulation, and Center of Overlap models are not mutually exclusive, but instead are operating in concert. In a recent review of paleontological data and species histories, Cowman and Bellwood conclude that the Coral Triangle had a long history of biodiversity accumulation stretching back to the Eocene, with subsequent diversification and contemporary export of species [22]. Briggs and Bowen reach a similar conclusion, that the Coral Triangle was a museum of biodiversity (accumulation) before it could be a cradle of speciation (export) [23]. A second dichotomy explored here is the competing hypotheses of speciation with or without geographic isolation. The emerging consensus is that both are important, and biodiversity may be highest when they work in concert, particularly for emerging species with adjacent (parapatric) ranges with alternative ecological regimes [4].

the yellow tang [64], lollysh [62], bullethead parrotsh [63], and ember parrotsh [60] have exported genetic and biological diversity, and export of biodiversity to the Coral Triangle has been shown for the lemon damselsh (Pomacentrus moluccensis) [78]. Rather than being an evolutionary graveyard, peripheral habitats can export biodiversity. The ecological component of speciation Although isolation can explain much of marine biodiversity, there is growing recognition that speciation in the sea can also occur in the presence of gene ow. The consensus is that this occurs when divergent selection overwhelms the homogenizing effect of gene ow [79,80]. New species can arise through intense competition in biodiversity hotspots, and these may constitute the majority of novel evolutionary lineages. However, new species are also evolving under conditions of ecological release in the depauperate archipelagos that surround biodiversity hotspots. In the Coral Triangle, a carnivorous grouper (family Epinephelidae) might overlap with over 100 other grouper species, in particular 15 members of the widespread genus Cephalopholis with similar ecology and diet [81]. Under such conditions, there is a premium on habitat partitioning and specialization that leads to rapid speciation, with subsequent radiation of the tter species into new areas. In the peripheral archipelago of Hawaii, an introduced grouper (C. argus) has to contend with no other members of the genus and only one other member of the taxonomic family [82], so that ecological release has allowed expansion of niche breadth [83]. This pattern is consistent with the ecological opportunity hypothesis proposed by Simpson [84] and recently supported by empirical studies [85,86]. The intense competition in the center, in conjunction with ecological release at the periphery, results in biodiversity feedback that produces more species than either process could alone.
6

Marine versus terrestrial speciation It remains to be seen whether biodiversity feedback is operating in terrestrial hotspots, but our focus remains on the tropical reef ecosystems that are the heart of marine biodiversity. In a departure from our illustrious predecessors, new research indicates that speciation has a different blend of processes in the sea. As noted by Dawson and Hamner [87], there is no hard dichotomy between speciation mechanisms in marine and terrestrial spheres; diversication is proceeding along both geographic and ecological partitions. However, the greater dispersal ability of most marine organisms reduces the prevalence of the former and tips the balance towards the latter. This contrast is clearly illustrated in the natural history of the Hawaiian Islands. Above the waterline, hundreds of terrestrial species have radiated from single rare colonization events. Examples range from owering plants to insects to birds, and the footprints of allopatric speciation are strong [88]. Below the waterline, however, allopatric divergence is apparent in low-dispersal species, but radiations within the archipelago have a stronger ecological component [8991]. Another important distinction is that Hawaii exports marine biodiversity, whereas exports of terrestrial diversity are virtually unknown. Hence, it is apparent that the origins and maintenance of biodiversity are starkly different in the sea, as is the prevalence of various mechanisms of speciation. Concluding remarks Two decades of phylogeographic research indicate that both allopatric and sympatric speciation is occurring in marine ecosystems, with the combination of ecological divergence and partial isolation (parapatry) perhaps offering the richest opportunities for diversication. Marine biodiversity hotspots are wellsprings that export species, but isolated archipelagos, marginal seas, and depauperate coastlines also contribute to overall biodiversity. Furthermore, the interaction between hotspots and peripheral habitats can boost biodiversity in both. The Indo-Pacic hotspot with the greatest diversity has a halo of archipelagos in both the Indian Ocean and (especially) the Pacic Ocean. The Caribbean hotspot has no halo, but has demonstrated feedback with depauperate Brazilian reefs [13]. Under this biodiversity feedback model, the islands that surround the central Indo-Pacic Coral Triangle are essential to the processes that support the highest marine biodiversity on the planet. Although phylogeographic and phylogenetic studies set the stage for this paradigm shift, the eld is entering a vast new stage of exploration. It is now possible to read entire genomes and divide genomes into arrays of neutral loci and regions under selection [58]. In abundant cases of suspected speciation by natural selection in sympatry (versus allopatry and secondary contact), genomics will reveal which loci are under selection and whether these are responsible for evolutionary divergences that are the foundation of biodiversity. New coalescent modeling approaches hold promise for revealing the timing of the cessation of gene ow and the relative roles of drift and selection responsible for evolutionary divergences [92]. It is certain that paradigms will fall, new ones will rise, and science will take a giant step towards the resolution of

TREE-1666; No. of Pages 8

Review
Darwins mystery. Furthermore, it is apparent that if human minds can conjure a plausible speciation mechanism, then nature can probably provide an example, and perhaps more as yet undreamed.
Acknowledgments
This synthesis greatly beneted from discussions with J.C. Briggs, F.W. Allendorf, K.R. Andrews, J.C. Avise, P.H. Barber, D.R. Bellwood, G. Bernardi, J.H. Choat, J. Copus, I. Fernandez-Silva, Z. Forsman, W.S. Grant, P. Jokiel, R. Kinzie, H. Lessios, G. Paulay, R. Pyle, J.E. Randall, C. Rocha, J. Roughgarden, and Z. Szabo. We thank the director and staff of the Hawaii Institute of Marine Biology, R. Kosaki, D. Pence, and members of the ToBo Laboratory for logistic assistance. For photographs we thank P.L. Munday, R. Myers, K.R. Nicastro, J.R. Pawlik, G. Paulay, J.E. Randall, K. Ruetzler, J. Sprung, B. Victor, and R. Williams. We also thank Editor P. Craze, H. Choat, and an anonymous reviewer for valuable suggestions and guidance. Studies of Indo-Pacic reef fauna by the ToBo Laboratory were supported by the National Science Foundation (OCE0453167 and OCE- 0929031 to B.W.B., OCE-0623678 to R.J.T.), the University of Hawaii Sea Grant College Program, the Seaver Institute (to B.W.B), the National Geographic Society (to M.T.C. and J.D.D.), the NOAA Ofce of National Marine Sanctuaries (to R.J.T.), and the California Academy of Sciences (to L.A.R.). This is HIMB publication 1544 and SOEST publication 8878.

Trends in Ecology & Evolution xxx xxxx, Vol. xxx, No. x

References
1 Mayr, E. (1954) Geographic speciation in tropical echinoids. Evolution 8, 118 2 Mayr, E. (1942) Systematics and the Origin of Species, From the Viewpoint of a Zoologist, Columbia University Press 3 Nelson, G. and Rosen, D.E., eds (1981) Vicariance Biogeography, A Critique, Columbia University Press 4 Rocha, L.A. and Bowen, B.W. (2008) Speciation in coral reef shes. J. Fish Biol. 72, 11011121 5 Spalding, M. and Grenfell, A.M. (1997) New estimates of global and regional coral reef areas. Coral Reefs 16, 225230 6 Almany, G.R. et al. (2007) Local replenishment of coral reef sh populations in a marine reserve. Science 316, 742744 7 Eble, J.A. et al. (2011) Not all larvae stay close to home: long-distance dispersal in Indo-Pacic reef shes, with a focus on the brown surgeonsh (Acanthurus nigrofuscus). J. Mar. Biol. 2011, 518516 8 Kinlan, B.P. et al. (2005) Propagule dispersal and the scales of marine dispersal. Divers. Distrib. 11, 139148 9 Kinlan, B.P. and Gaines, S.D. (2003) Propagule dispersal in marine and terrestrial environments: a community perspective. Ecology 84, 20072020 10 Reece, J.S. et al. (2011) Long larval duration in moray eels (Muraenidae) ensures ocean-wide connectivity despite differences in adult niche breadth. Mar. Ecol. Prog. Ser. 437, 269277 11 Horne, J.B. et al. (2008) High population connectivity across the IndoPacic: congruent lack of phylogeographic structure in three reef sh congeners. Mol. Phylogenet. Evol. 49, 629638 12 DiBattista, J.D. et al. (2012) Twisted sister species of pygmy angelshes: discordance between taxonomy, coloration, and phylogenetics. Coral Reefs 31, 839851 13 Rocha, L.A. et al. (2008) Comparative phylogeography of Atlantic reef shes indicates both origin and accumulation of diversity in the Caribbean. BMC Evol. Biol. 8, 157 14 Briggs, J.C. (2003) Marine centres of origin as evolutionary engines. J. Biogeogr. 30, 118 15 Barber, P.H. et al. (2011) Evolution and conservation of marine biodiversity in the Coral Triangle: insights from stomatopod Crustacea. Crustacean Issues 19, 129156 16 Timm, J. et al. (2008) Contrasting patterns in species boundaries and evolution of anemoneshes (Amphiprioninae, Pomacentridae) in the centre of marine biodiversity. Mol. Phylogenet. Evol. 49, 268276 17 Carpenter, K.E. et al. (2011) Comparative phylogeography of the Coral Triangle and implications for marine management. J. Mar. Biol. 2011, 396982 18 Mora, C. et al. (2003) Patterns and processes in reef sh diversity. Nature 421, 933936

19 Craig, M.T. et al. (2010) Origins, ages, and populations histories: comparative phylogeography of endemic Hawaiian butteryshes (genus Chaetodon). J. Biogeogr. 37, 21252136 20 Daly-Engel, T.S. et al. (2012) Global phylogeography with mixed marker analysis reveals male-mediated dispersal in the endangered scalloped hammerhead shark (Sphyrna lewini). PLoS ONE 7, e29986 21 Whitney, N.M. et al. (2012) Phylogeography of the whitetip reef shark (Triaenodon obesus): a sedentary species with a broad distribution. J. Biogeogr. 39, 11441156 22 Cowman, P.F. and Bellwood, D.R. (2013) The historical biogeography of coral reef shes: global patterns of origination and dispersal. J. Biogeogr. 40, 209224 23 Briggs, J.C. and Bowen B.W. Evolutionary patterns: marine shelf habitat. J. Biogeogr. http://dx.doi.org/10.1111/jbi.12082 24 Hourigan, T.F. and Reese, E.S. (1987) Mid-ocean isolation and the evolution of Hawaiian reef shes. Trends Ecol. Evol. 2, 187191 25 Jokiel, P. (1987) Ecology, biogeography and evolution of corals in Hawaii. Trends Ecol. Evol. 2, 179182 26 Kay, E.A. and Palumbi, S.R. (1987) Endemism and evolution in Hawaiian marine invertebrates. Trends Ecol. Evol. 2, 183187 27 Jokiel, P. and Martinelli, F.J. (1992) The vortex model of coral reef biogeography. J. Biogeogr. 19, 449458 28 Connolly, S.R. et al. (2003) Indo-Pacic biodiversity of coral reefs: deviations from a mid-domain model. Ecology 84, 21782190 29 Budd, A.F. and Pandol, J.M. (2010) Evolutionary novelty is concentrated at the edge of coral species distributions. Science 328, 15581561 30 Hobbes, J-P.A. et al. (2009) Marine hybrid hotspot at Indo-Pacic biogeographic border. Biol. Lett. 5, 258261 31 Gaither, M.R. et al. (2011) Phylogeography of the reef sh Cephalopholus argus (Epinephelidae) indicates Pleistocene isolation across the Indo-Pacic Barrier with contemporary overlap in the Coral Triangle. BMC Evol. Biol. 11, 189 32 Bernardi, G. et al. (2004) Evolution of coral reef sh Thalassoma spp. (Labridae): 1. Molecular phylogeny and biogeography. Mar. Biol. 144, 369375 33 Hodge, J.R. et al. (2011) The role of peripheral endemism in species diversication: evidence from the coral reef sh genus Anampses (Family Labridae). Mol. Phylogenet. Evol. 62, 653663 34 Dobzhansky, T. (1937) Genetics and the Origin of Species, Columbia University Press 35 Coyne, J.A. and Orr, H.A. (2004) Speciation, Sinauer Associates 36 Crow, K.D. et al. (2010) Sympatric speciation in a genus of marine reef shes. Mol. Ecol. 19, 20892105 37 Quesada, H. et al. (2007) Phylogenetic evidence for multiple sympatric ecological diversication in a marine snail. Evolution 61, 16001612 38 Duran, S. and Rutzler, K. (2006) Ecological speciation in a Caribbean marine sponge. Mol. Phylogenet. Evol. 40, 292297 39 Puebla, O. et al. (2007) Colour pattern as a single trait driving speciation in Hypoplectrus coral reef shes? Proc. R. Soc. Lond. B: Biol. Sci. 274, 12651271 40 Williams, S.T. (2000) Species boundaries in the starsh genus Linckia. Mar. Biol. 136, 137148 41 Taylor, M.S. and Hellberg, M.E. (2005) Marine radiations at small geographic scales: speciation in Neotropical reef gobies (Elacatinus). Evolution 59, 374385 42 Leray, M. et al. (2010) Allopatric divergence and speciation in coral reef sh: the three-spot dascyllus, Dascyllus trimaculatus, species complex. Evolution 6465, 12181230 43 Morin, P.A. et al. (2010) Complete mitochondrial genome phylogeographic analysis of killer whales (Orcinus orca) indicates multiple species. Genome Res. 20, 908916 44 Andrews, K.R., et al. The evolving male: spinner dolphin (Stenella longirostris) ecotypes are divergent at Y chromosome but not mtDNA or autosomal markers. Mol. Ecol. (in press) 45 Corrigan, S. and Beheregaray, L.B. (2009) A recent shark radiation: molecular phylogeny, biogeography and speciation of wobbegong sharks (family: Orectolobidae). Mol. Phylogenet. Evol. 52, 205216 46 Kashiwagi, T. et al. (2012) The genetic signature of recent speciation in manta ray (Manta alfredi and M. birostris). Mol. Phylogenet. Evol. 64, 212218
7

TREE-1666; No. of Pages 8

Review
47 Choat, H. (2006) Phylogeography and reef shes: bringing ecology back into the argument. J. Biogeogr. 33, 967968 48 Doebeli, M. and Dieckmann, U. (2003) Speciation along environmental gradients. Nature 421, 259264 49 Quenouille, B. et al. (2011) Speciation in tropical seas: allopatry followed by range change. Mol. Phylogenet. Evol. 58, 546552 50 Rocha, L.A. et al. (2008) Historical biogeography and speciation in the reef sh genus Haemulon (Teleostei: Haemulidae). Mol. Phylogenet. Evol. 48, 918928 51 Gavrilets, S. (2000) Waiting time to parapatric speciation. Proc. R. Soc. Lond. B: Biol. Sci. 267, 24832492 52 Hellberg, M.E. and Vaquier, V.D. (1999) Rapid evolution of fertilization selectively and lysine cDNA sequences in teguline gastropods. Mol. Biol. Evol. 16, 839848 53 Hendry, A.P. et al. (2000) Rapid evolution of reproductive isolation in the wild. Science 290, 516518 54 Nosil, P. and Schluter, D. (2011) The genes underlying the process of speciation. Trends Ecol. Evol. 26, 160167 55 Duda, T.F. and Rolan, E. (2005) Explosive radiation of Cape Verde Conus: a marine species ock. Mol. Ecol. 14, 267272 56 Puritz, J.B. et al. (2012) Extraordinarily rapid life history divergence between Cryptasterina sea star species. Proc. R. Soc. Lond. B: Biol. Sci. 279, 39143922 57 Ruber, L. and Zardoya, R. (2005) Rapid cladogenesis in marine shes revisited. Evolution 59, 11191127 58 Van Doorn, G.S. et al. (2009) On the origin of species by natural and sexual selection. Science 326, 17041707 59 Bellemain, E. and Ricklefs, R.E. (2008) Are islands the end of the colonization road? Trends Ecol. Evol. 23, 461468 60 Fitzpatrick, J.M. et al. (2011) The West Pacic diversity hotspot as a source or sink for new species? Population genetic insights from the Indo-Pacic parrotsh Scarus rubroviolaceus. Mol. Ecol. 20, 219234 61 Gaither, M.R. et al. (2011) High connectivity in the deepwater snapper Pristipomoides lamentosus (Lutjanidae) across the Indo-Pacic with isolation of the Hawaiian Archipelago. PLoS ONE 6, e28913 62 Skillings, D. et al. (2011) Gateways to Hawaii genetic population structure of the tropical sea cucumber Holothuria atra. J. Mar. Biol. 2011, 783030 63 Bay, L.K. et al. (2004) High genetic diversities and complex genetic structure in an Indo-Pacic tropical reef sh (Chlorurus sordidus): evidence of an unstable evolutionary past? Mar. Biol. 144, 757767 64 Eble, J.A. et al. (2011) Escaping paradise: larval export from Hawaii in an Indo-Pacic reef sh, the yellow tang (Zebrasoma avescens). Mar. Ecol. Prog. Ser. 428, 245258 65 DiBattista, J.D. et al. (2013) After continents divide: comparative phylogeography of reef shes from the Red Sea and Indian Ocean. J. Biogeogr. http://dx.doi.org/10.1111/jbi.12068 66 Choat, J.H. et al. (2012) Patterns and processes in the evolutionary history of parrotshes (family Labridae). Biol. J. Linn. Soc. 107, 529557 67 Robertson, D.R. et al. (2004) Tropical transpacic shore shes. Pac. Sci. 58, 507565 68 Lessios, H.A. and Robertson, D.R. (2006) Crossing the impassable: genetic connections in 20 reef shes across the eastern Pacic barrier. Proc. R. Soc. Lond. B: Biol. Sci. 273, 22012208 69 Floeter, S.R. et al. (2008) Atlantic reef sh biogeography and evolution. J. Biogeogr. 35, 2247 70 Liao, P.C. et al. (2007) Phylogeography of Ceriops tagal (Rhizophoraceae) in Southeast Asia: the land barrier of the Malay Peninsula has caused population differentiation between the Indian Ocean and South China Sea. Conserv. Genet. 8, 8998 71 Williams, S.T. (2007) Origins and diversication of Indo-West Pacic marine fauna: evolutionary history and biogeography of turban shells (Gastropoda, Turbinidae). Biol. J. Linn. Soc. 92, 573592 72 Teske, P.R. et al. (2005) Molecular evidence for long distance colonization in an Indo-Pacic seahorse lineage. Mar. Ecol. Prog. Ser. 286, 249260 73 Halas, D. and Winterbottom, R. (2009) A phylogenetic test of multiple proposals for the origins of the East Indies coral biota. J. Biogeogr. 36, 18471860 74 Barber, P.H. and Bellwood, D.R. (2005) Biodiversity hotspots: evolutionary origins of biodiversity in wrasses (Halichoeres: Labridae) in the Indo-Pacic and New World tropics. Mol. Phylogenet. Evol. 35, 235253
8

Trends in Ecology & Evolution xxx xxxx, Vol. xxx, No. x

75 Lessios, H.A. (2008) The great American schism: divergence of marine organisms after the rise of the Central American Isthmus. Annu. Rev. Ecol. Evol. Syst. 39, 6391 76 Meyer, C.P. (2003) Molecular systematics of cowries (Gastropoda: Cypraeidae) and diversication patterns in the tropics. Biol. J. Linn. Soc. 79, 401459 77 Malay, M.C.D. and Paulay, G. (2009) Peripatric speciation drives diversication and distribution patterns of reef hermit crabs (Decapoda: Diogenidae: Calcinus). Evolution 64, 634662 78 Drew, J. and Barber, P.H. (2009) Sequential cladogenesis of the reef sh Pomacentrus moluccensis (Pomacentridae) supports the peripheral origin of marine biodiversity in the Indo-Australian archipelago. Mol. Phylogenet. Evol. 53, 336339 79 Bolnick, D.I. and Fitzpatrick, B.M. (2007) Sympatric speciation: models and empirical evidence. Annu. Rev. Ecol. Evol. Syst. 38, 459487 80 Crow, K.D. et al. (2007) Maintenance of species boundaries despite rampant hybridization between three species of reef shes (Hexagrammidae): implications for the role of selection. Biol. J. Linn. Soc. 91, 135147 81 Allen, G.R. and Adrim, M. (2003) Coral reef shes of Indonesia. Zool. Stud. 42, 172 82 Mundy, B.C. (2005) Checklist of the Fishes of the Hawaiian Archipelago (Bishop Museum Bulletin in Zoology, Vol. 6), Bishop Museum Press 83 Meyer, A.L. and Dierking, J. (2011) Elevated size and body condition and altered feeding ecology of the grouper Cephalopholis argus in nonnative habitats. Mar. Ecol. Prog. Ser. 439, 203212 84 Simpson, G.G. (1953) The Major Features of Evolution, Columbia University Press 85 Duda, T.F. and Lee, T. (2009) Ecological release and the venom evolution of a predatory snail at Easter Island. PLoS ONE 4, e5558 86 Yoder, J.B. et al. (2010) Ecological opportunity and the origin of adaptive radiations. J. Evol. Biol. 23, 15811596 87 Dawson, M.N. and Hamner, W.M. (2008) A biophysical perspective on dispersal and the geography of evolution in marine and terrestrial systems. J. R. Soc. Interface 5, 135150 88 Wagner, W.L. and Funk, V.A. (1995) Hawaiian Biogeography: Evolution on a Hot Spot Archipelago, Smithsonian Institution Press 89 Faucci, A. et al. (2007) Host shift and speciation in a coral-feeding nudibranch. Proc. R. Soc. Lond. B: Biol. Sci. 274, 111119 90 Forsman, Z.H. et al. (2010) Ecomorph or endangered coral? DNA and microstructure reveal Hawaiian species complexes: Montipora dilatata/abellata/turgescens & M. patula/verrilli. PLoS ONE 5, e15021 91 Bird, C.E. et al. (2011) Diversication of endemic sympatric limpets (Cellana spp.) in the Hawaiian Archipelago. Mol. Ecol. 20, 21282141 92 Ilves, K.L. et al. (2010) Colonization and/or mitochondrial selective sweeps across the North Atlantic intertidal assemblage revealed by multi-taxa approximate Bayesian computation. Mol. Ecol. 19, 45054519 93 Briggs, J.C. and Bowen, B.W. (2012) A realignment of marine biogeographic provinces with particular reference to sh distributions. J. Biogeogr. 39, 1230 94 Zardi, G.I. et al. (2011) Adaptive traits are maintained on steep selective gradients despite gene ow and hybridization in the intertidal zone. PLoS ONE 6, e19402 95 Rocha, L.A. et al. (2005) Ecological speciation in tropical reef shes. Proc. R. Soc. Lond. B: Biol. Sci. 272, 573579 96 Ru tzler, K. et al. (2007) Adaptation of reef and mangrove sponges to stress: evidence for ecological speciation exemplied by Chondrilla caribensis new species (Demospongiae, Chondrosida). Mar. Ecol. 28S1, 95111 97 Victor, B.C. and Randall, J.E. (2010) Gramma dejongi, a new basslet (Perciformes: Grammatidae) from Cuba, a sympatric sibling species of G. loreto. Zool. Stud. 49, 865871 98 Bongaerts, P. et al. (2010) Genetic divergence across habitats in the widespread coral Seriatopora hystrix and its associated Symbiodinium. PLoS ONE 5, e10871 99 Munday, P.L. et al. (2004) Evidence for sympatric speciation by host shift in the sea. Curr. Biol. 14, 14981504 100 Mouritsen, H. and Ritz, T. (2005) Magnetoreception and its use in bird navigation. Curr. Opin. Neurobiol. 15, 406414