Comparative phylogeography of endemic Azorean arthropods

For a remote oceanic archipelago of up to 8 Myr age, the Azores have a comparatively low level of endemism. We present an analysis of phylogeographic patterns of endemic Azorean island arthropods aimed at testing patterns of diversification in relation to the ontogeny of the archipelago, in order to distinguish between alternative models of evolutionary dynamics on islands. We collected individuals of six species (representing Araneae, Hemiptera and Coleoptera) from 16 forest fragments from 7 islands. Using three mtDNA markers, we analysed the distribution of genetic diversity within and between islands, inferred the differentiation time-frames and investigated the inter-island migration routes and colonization patterns.

Each species exhibited very low levels of mtDNA divergence, both within and between islands. The two oldest islands were not strongly involved in the diffusion of genetic diversity within the archipelago. The most haplotype-rich islands varied according to species but the younger, central islands contributed the most to haplotype diversity. Colonization events both in concordance with and in contradiction to an inter-island progression rule were inferred, while a non-intuitive pattern of colonization from western to eastern islands was also inferred.

The geological development of the Azores has followed a less tidy progression compared to classic hotspot archipelagos, and this is reflected in our findings. The study species appear to have been differentiating within the Azores for <2 Myr, a fraction of the apparent life span of the archipelago, which may indicate that extinction events linked to active volcanism have played an important role. Assuming that after each extinction event, colonization was initiated from a nearby island hosting derived haplotypes, the apparent age of species diversification in the archipelago would be moved closer to the present after each extinction-recolonization cycle. Exploiting these ideas, we propose a general model for future testing.