Aim The main Japanese islands are land-bridge islands divided by the biogeographic division Blakiston’s Line and represent two natural laboratories for studying land-bridge diversification. Colonization of the current mammal fauna has been dated to the middle to late Pleistocene using fossil evidence. The purpose of this paper is to apply a molecular clock to the genetic divergences between Japanese mammalian taxa and their sister mainland taxa to test the late Pleistocene land-bridge colonization hypothesis.
Location The main Japanese islands (Kyushu, Shikoku, Honshu and Hokkaido).
Methods I used mitochondrial DNA (cytochrome b) and a species tree approach to estimate the divergence times of 24 Japanese non-volant terrestrial mammal taxa and their mainland sister taxa using the program *beast. I then tested for evidence of non-simultaneous divergence among these taxon-pairs by controlling for expected coalescent stochasticity using the program MsBayes.
Results Divergence events between taxa on Japan and their mainland sister taxa were significantly older than expected under the current paradigm, which is based on fossil data. Consistent with the land-bridge colonization hypothesis, there was evidence of multiple divergence events.
Main conclusions These results implicate a colonization timeframe that is older than posited by the current paradigm based on fossil evidence. However, these results are still consistent with the land-bridge colonization hypothesis. Multiple periods of land-bridge connectivity may account for the current mammalian fauna in Japan. In addition, half of the divergence time estimates in the Honshu–Shikoku–Kyushu region were clumped around 2.4 Ma, which might suggest a dramatic interchange period, concordant with a period of significant global cooling, when the first land bridge may have connected Japan to the mainland.