Amphibians are thought to be unable to disperse over ocean barriers because they do not tolerate the osmotic stress of salt water. Their distribution patterns have therefore generally been explained by vicariance biogeography. Here, we present compelling evidence for overseas dispersal of frogs in the Indian Ocean region based on the discovery of two endemic species on Mayotte. This island belongs to the Comoro archipelago, which is entirely volcanic and surrounded by sea depths of more than 3500 m. This constitutes the first observation of endemic amphibians on oceanic islands that did not have any past physical contact to other land masses. The two species of frogs had previously been thought to be nonendemic and introduced from Madagascar, but clearly represent new species based on their morphological and genetic differentiation. They belong to the genera Mantidactylus and Boophis in the family Mantellidae that is otherwise restricted to Madagascar, and are distinguished by morphology and mitochondrial and nuclear DNA sequences from mantellid species occurring in Madagascar. This discovery permits us to update and test molecular clocks for frogs distributed in this region. The new calibrations are in agreement with previous rate estimates and indicate two further Cenozoic transmarine dispersal events that had previously been interpreted as vicariance: hyperoliid frogs from Africa to Madagascar (Heterixalus) and from Madagascar to the Seychelles islands (Tachycnemis). Our results provide the strongest evidence so far that overseas dispersal of amphibians exists and is no rare exception, although vicariance certainly retains much of its importance in explaining amphibian biogeography.

Poison frogs of the family Dendrobatidae contain cryptic as well as brightly colored, presumably aposematic species. The prevailing phylogenetic hypothesis assumes that the aposematic taxa form a monophyletic group while the cryptic species (Colostethus sensu lato) are basal and paraphyletic. Analysis of 86 dendrobatid sequences of a fragment of the 16S rRNA gene resulted in a much more complex scenario, with several clades that contained aposematic as well as cryptic taxa. Monophyly of the aposematic taxa was significantly rejected by SH-tests in an analysis with additional 12S and 16S rDNA fragments and reduced taxon sampling. The brightly colored Allobates femoralis and A. zaparo (Silverstone) comb. nov. (previously Epipedobates) belong in a clade with cryptic species of Colostethus. Additionally, Colostethus pratti was grouped with Epipedobates, and Colostethus bocagei with Cryptophyllobates. In several cases, the aposematic species have general distributions similar to those of their non-aposematic sister groups, indicating multiple instances of regional radiations in which some taxa independently acquired bright color. From a classificatory point of view, it is relevant that the type species of Minyobates, M. steyermarki, resulted as the sister group of the genus Dendrobates, and that species of Mannophryne and Nephelobates formed monophyletic clades, corroborating the validity of these genera. Leptodactylids of the genera Hylodes and Crossodactylus were not unambiguously identified as the sister group of the Dendrobatidae; these were monophyletic in all analyses and probably originated early in the radiation of Neotropical hyloid frogs.

No abstract

Aim The biogeographical origins of the extant vertebrates endemic to Madagascar are largely unsolved, but have often been related to vicariance in the context of fragmentation of the supercontinent Gondwana in the Mesozoic. Such hypotheses are especially appealing in the case of cichlid ®shes, which show phylogenetic relationships re¯ecting the temporal successions of the breakup of Gondwana. We used molecular clock data to test this assumption.Location Fragments of the 16S rRNA gene and of the nuclear Tmo-4C4 locus, partly obtained from Genbank from South American, African, Malagasy and Indian cichlids were analysed.Methods Based on monophyletic cichlid radiations in African lakes, we calibrated a molecular clock. The obtained rates were used to estimate the age of divergence of the major cichlid clades. ResultsThe results agreed better with a Cenozoic than with a Mesozoic divergence, and were in accordance with the fossil record. Sequence divergences of the 16S and 12S rRNA genes of most lineages of Malagasy terrestrial and freshwater vertebrates from their non-Malagasy sister groups were below saturation and many were relatively similar to those of cichlids.Main conclusions A Cenozoic dispersal from continental landmasses may explain the origin of most extant Malagasy vertebrate groups better than a Jurassic/Cretaceous vicariance.

No abstract