Phylogeography and reef fishes: bringing ecology back into the argument

Speciation in tropical marine fishes is an increasingly important topic. Analysis of the mechanisms responsible for the diversity of this assemblage and the actual identification of species is at its core. Molecular studies have contributed significantly to our understanding of the processes responsible for the complex patterns of genetic diversification in tropical seas (Grant & Bowen, 1998) and are central to our understanding of reef biodiversity. However, our capacity to resolve issues relevant to the study of marine speciation faces numerous challenges. Reef fishes are a complex group comprising numerous lineages with distinctive life-history features. These consist of complexes of co-occurring species, the majority of which have broad geographical distributions and larvae capable of wide dispersal. Fragmented habitats are connected by oceanic waters that lack obvious barriers to dispersal. Moreover, the environmental history of coral reefs is exceedingly complex. Most coral reef organisms are restricted to very shallow depths and thus are subject to major episodes of habitat disturbance, including sea level and temperature fluctuations over geologically short time periods. The demographic history of reef fishes is predictably one of population expansion and contraction with episodes of secondary contact and hybridization, all of which can result in populations characterized by non-equilibrium genetic structures (Fauvelot et al., 2003). One consequence of this heritage is the evident difficulties faced in interpreting the phylogeography of reef fish species. Geographic patterns of genetic structure vary substantially among different groups. To date, much of this information is based on mitochondrial gene fragments. In some instances, examination of additional markers, including those from the nuclear genome, has clarified the mechanisms responsible for genetic structure (e.g. Lemaire et al., 2000– although these studies focus on temperate waters). However, the task requires more than just an updating of the molecular tool kit. It is now time to review many of the approaches that drive our investigation of the evolutionary biology of reef fishes and bring information on post-settlement processes, life histories and demography to bear on the issue. Moreover, the critical importance of well-resolved taxonomies must be emphasized and integrated into the study of evolutionary processes. The integration of ecological and evolutionary processes must acknowledge the differing time scales and rates operating in each case. Ecological connectivity among fragmented populations requires sufficient input of propagules to modify demographic processes such as recruitment. Genetic connectivity does not imply persistent demographic exchange. Very low levels of migration are sufficient to homogenize population structure. The ecological and evolutionary processes interact in interesting ways in reef fish populations. Demographic differentiation can occur in adjacent populations that are apparently genetically homogenous and subject to substantial genetic exchange, suggesting an important role for phenotypic plasticity in reef fish populations. In contrast, populations of some species may be genetically differentiated at very local scales (see below). The most compelling cases involve populations that have the capacity to exchange propagules between adult habitats separated by thousands of kilometres, yet appear to be unable to cross relatively small oceanic distances between adjacent reef systems. The possible action of ecological processes in inhibiting the establishment of migrants, even when populations are adjacent, is a potentially rewarding area of investigation. To date, interpretation of phylogeographic structure in tropical marine organisms has been heavily skewed towards explanations based on attributes of the larvae and their pelagic environment, including pelagic larval durations (PLD), larval behaviour, and multiscale patterns of oceanic circulation (Swearer et al., 2002). While it is known that wide dispersal can be achieved in reef species (Bowen et al., 2001), attention has recently focused on examples of restricted larval dispersal and finer scale population structure than might have been predicted given the lack of geographic isolation in ocean environments and possession of dispersive propagules (Swearer et al., 2002). However, the larval journey itself may not be the problem. Studies of reef fishes have provided numerous examples of the importance of post-settlement processes in structuring populations (Hixon & Jones, 2005). In the absence of plausible mechanisms inhibiting dispersal and migration, alternative explanations based on ecological processes operating at the time of settlement must be considered (Rocha et al., 2005). We might therefore ask, to what extent are ecological processes acting on the settlement and early recruitment stages of reef fishes responsible for patterns of genetic differentiation observed in some populations? The apparent barriers to dispersal often represent boundaries between insular and continental populations and suggest a role for early habitat selection and recruitment. In this context field-based measures of selection strength could be important in determining whether genetic divergence could have occurred in the face of ongoing migration. The study by Craig et al. (2006) in this issue, considers the phylogeography of two closely related species of Epinephelus, the dominant genus of tropical groupers, and illustrates a number of the issues discussed above. The focus is the role of biogeographic boundaries in shaping the relationships of the species within the tropical East Pacific. One species has a geographic distribution restricted to offshore islands, while the second is more cosmopolitan, occupying an extensive coastal area of the east Pacific and some offshore islands. Given the distance involved, the possibility of appreciable levels of gene flow amongst all populations is a reasonable expectation. In attempting to evaluate the status of the two putative sibling species, three hypotheses were considered: (1) the populations will be confirmed as species that persist in the face of ongoing gene flow; (2) they represent species that are in secondary contact resulting in hybridization; and (3) the populations reflect recent speciation but with incomplete lineage sorting. A 468 bp fragment of the mitochondrial gene cyt b provided the basis for phylogeographic analysis. The study revealed restricted gene flow between the insular and coastal habitats of the two species, but with no evidence of reciprocal monophyly. The authors’ conclusion was recent speciation, as indicated by the distinctive colour patterns, but with incomplete lineage sorting. However, while the mitochondrial marker did provide evidence of restricted gene flow, there were a number of possible additional explanatory scenarios, including hybridization among the two taxa. The nature of these findings is indicative of the complex nature of evolutionary divergence in tropical fishes, especially in the context of colour morphs. An intriguing finding, which mirrored a similar study in the tropical Atlantic (Carlin et al., 2003), was that populations could be partitioned genetically into (1) cosmopolitan elements that appeared to transcend oceanic and biogeographic barriers and (2) spatially discrete populations restricted to islands or small areas of coasts. In both studies, adjacent populations showed restricted gene flow while more distant populations exchanged propagules. Both studies concluded that patterns of pelagic larval duration alone were uninformative. The restriction of gene flow across narrow boundaries may indicate the importance of vicariant barriers such as current systems. Restriction of haplotypes to particular locations, such as Clipperton Island, may indicate larval retention mechanisms. The problem that arises with explanations based on patterns of larval dispersal is that there is so much variation amongst species and lineages in their phylogeographic structure. Does this always reflect variation in their capacity to disperse? To reiterate, although investigations to date have emphasized larval dispersal patterns almost exclusively, the journey itself may not be the problem; the crucial issues could involve the processes of arrival and settling in. Strong habitat-mediated selection acting on recruits that differ in behavioural, morphological and demographic features could also determine genetic structure. Ecological investigations of recruits linked to genetic analyses and comparative analyses of age-based dynamics are a priority. Distinguishing the role of processes reflecting dispersal and larval features from those associated with post-settlement events, in populations that are unlikely to be in equilibrium in a genetic context, is a challenging focus for future investigation. Demographic analyses are important for an additional reason. Many studies now make use of coalescent analysis in which the estimation of generation time is a critical variable. As many reef fishes now appear to be long-lived, with extended reproductive lives (Choat & Robertson, 2002), age at first maturity may not be an appropriate indication of generation time. Moreover, long generation times and overlapping generations may extend the time taken to establish genetic equilibrium. An additional concern relates to taxonomic practice. Analyses of mitochondrial genomes in reef fishes have made it clear that fragments of a single gene may not be sufficient to resolve species identity. This suggests that attempts using the single mt fragment CO1 to ‘barcode’ reef fishes should be treated with scepticism, especially if species discovery is a goal. Craig et al. (2006) argue that it is imperative to retain voucher specimens in museums so that identifications may be confirmed. The bad news is that with many species-rich but morphologically conservative reef fish taxa, preserved specimens may not be readily identified. There is no substitute for authoritative field identifications, especially where the goal is collection of tissue samples for molecular analysis. However, this requires the input of taxonomic expertise: something that is in rapid decline. An informed approach to the evolution of reef fishes requires not only familiarity with molecular tools but knowledge of the morphological and ecological criteria by which species are distinguished.