Molecular genetics of cirrhitoid fishes (Perciformes: Cirrhitoidea): phylogeny, taxonomy, biogeography, and stock structure

The molecular phylogenetic relationships within three of the five cirrhitoid fish families were reconstructed from mitochondrial DNA cytochrome b, cytochrome oxidase I, and D-loop sequences. Analysis of the Cheilodactylidae provided evidence that much taxonomic revision is required. The molecular data suggest that this family should be restricted to the two South African Cheilodactylus, as they are highly divergent from the other cheilodactylids and one member is the type species. The remaining 25 cheilodactylids should be transferred to the Latridae. Nine of the non South African Cheilodactylus can be allocated to three new genera; Goniistius (elevated to generic rank), Zeodrius (resurrected), and Morwong (resurrected), while the placement of three species is uncertain. The three South African Chirodactylus should revert to Palunolepis, as they are distinct from the South American type species Chirodact_vlus variegatus. Acantholatris clusters within Nemadactylus, ar;d the former should be synonymised. Cryptic speciation has occurred within Cheilodactylus (Goniistius) vittatus. The generic allocation of the four latrid species is sound, although this family should be expanded to encompass all but two cheilodactylids. Relative levels of genetic divergence within the Aplodactylidae support the most recent revision of this family, during which the monotypic genus Crinodus was synonymised with Aplodactylus. Mokcular phylogenetic relationships and estimates of divergence time obtained from molecular clock calibrations suggest a dominant role of long distance dispersal for the present distribution of cheilodactylid and aplodactylid fishes. Suggestions that ancestral taxa were vicariantly isolated during the fragmentation of Gondwana are rejected, as estimated divergence times appreciably postdate this event. Dispersal and radiation of Nemadactylus and Acantholatris throughout the Southern Ocean was particularly recent, occurring within the last 0.6-2.6 Myr. The waters of Australia and New Zealand represent a likely origin for this dispersal, and at least two events are identified, one eastward. Similarly, it appears that aplodactylids also originated in the waters of Australia and New Zealand, but in this instance the majority of radiation was undertaken prior to colonisation of the southeastern Pacific. Ocean currents and long duration offshore pelagic larvae probably facilitated dispersal. Phylogeographic analysis of the antitropically-distributed cheilodactylid subgenus Goniistius identified three transequatorial divergences, rather than a minimum of two as inferred from the distributions of individual taxa. The identified divergences also occurred during two distinct periods, the mid Miocene and mid to late Pliocene, and are best explained by chance dispersal or vicariance resulting from biotic interactions or temperature changes. The levels of genetic separation for three cirrhitoid species pairs with east-west allopatric distributions across southern Australia reject the possibility that the members of each pair diverged simultaneously during a shared vicariance event. Although the levels of genetic separation were similar for Goniistius and Aplodactylus pairs, separate north and south coast \ vicariance events are invoked based on likely thermal tolerances. Speciation resulting from chance dispersal and the founding of new populations is rejected due to the absence of barriers sufficiently large to isolate taxa with such high dispersal capabilities. Estimated divergence times fall between the late Miocene and mid Pliocene and fail to implicate recent · Pleistocene glaciations. Seven microsatellite loci were characterised for Nemadactylus macropterus in an effort to resolve its stock structure in Australian waters and to assess the resolving power of different molecular techniques. Microsatellites did not identify any stock structuring in the waters of southern Australia. Divergence was also absent between Australian and New Zealand populations which contrasted the findings from allozyme and mitochondrial DNA studies. Homoplasy of alleles at highly polymorphic loci is offered as a possible explanation for the lower resolution of stock structure obtained with microsatellites.The microsatellites characterised for N. macropterus were also employed to examine the taxonomic status of the morphologically similar S0uth American species N. bergi. Separate status was supported by divergence at a single locus. Microsatellites also provided evidence for a recent bottleneck in the effective population size of N. bergi but not N. macropterus or A. monodactylus. Based on this observation the mitochondrial DNA lineage monophyly observed for N. hergi but not N. macropterus or A. monodactylus may reflect the influence of effective population size on the time required for complete sorting of mitochondrial lineages."