Island bat diets: does it matter more who you are or where you live?

Differences in body size, echolocation call frequency and location may result in diet partitioning among bat species. Comparisons between island populations are one way to evaluate these competing hypotheses. We conducted a species-level diet analysis of three Rhinolophus and one Hipposideros species on the Philippine islands of Cebu, Bohol and Siquijor. We identified 655 prey (MOTUs) in the guano from 77 individual bats. There was a high degree of overlap among species' diets despite differences in body size and call frequency. For example, the diet of the 3 g-Hipposideros pygmaeus (mean CF = 102 kHz) exhibited a diet overlap higher than expected by chance with all three Rhinolophus species, even the 13 g-Rhinolophus inops (mean CF = 54 kHz). We observed more convergence in diet between Rhinolophus species and H. pygmaeus than between Rhinolophus species themselves, which may be explained by the broad diet of H. pygmaeus. There was less dietary overlap between Rhinolophus virgo from two islands than between R. virgo and congeners from Cebu. These data suggest that location causes convergence in diet, but specific species characteristics may drive niche specialization. The complex interplay between location and the perceptual ability of each species leads to a situation where simple explanations, for example body size, do not translate into predictable prey partitioning. In particular, our observations raise interesting questions about the foraging strategy and adaptability of the tiny H. pygmaeus.

Single-locus species delimitation: a test of the mixed Yule-coalescent model, with an empirical application to Philippine round-leaf bats.

Prospects for a comprehensive inventory of global biodiversity would be greatly improved by automating methods of species delimitation. The general mixed Yule-coalescent (GMYC) was recently proposed as a potential means of increasing the rate of biodiversity exploration. We tested this method with simulated data and applied it to a group of poorly known bats (Hipposideros) from the Philippines. We then used echolocation call characteristics to evaluate the plausibility of species boundaries suggested by GMYC. In our simulations, GMYC performed relatively well (errors in estimated species diversity less than 25%) when the product of the haploid effective population size (N(e)) and speciation rate (SR; per lineage per million years) was less than or equal to 10(5), while interspecific variation in N(e) was twofold or less. However, at higher but also biologically relevant values of N(e) × SR and when N(e) varied tenfold among species, performance was very poor. GMYC analyses of mitochondrial DNA sequences from Philippine Hipposideros suggest actual diversity may be approximately twice the current estimate, and available echolocation call data are mostly consistent with GMYC delimitations. In conclusion, we consider the GMYC model useful under some conditions, but additional information on N(e), SR and/or corroboration from independent character data are needed to allow meaningful interpretation of results.

Philippine bats of the genus Kerivoula (Chiroptera: Vespertilionidae): Overview and assessment of variation in K. pellucida and K. whiteheadi.

Bats of the genus Kerivoula (Mammalia, Chiroptera) are widespread in the Philippines with four reported species, but have been poorly known due to a paucity of specimens. We provide the first molecular phylogeny for Philippine Kerivoula, which supports the existence of four distinct clades that we treat as species (K. hardwickii, K. papillosa, K. pellucida, and K. whiteheadi); these four overlap broadly geographically. Each of these may be recognized on the basis of cytochrome b sequences and external and craniodental morphology. Detailed examination of K. pellucida shows little geographic differentiation within the Philippines, but they differ subtly from those on the Sunda Shelf. We consider K. whiteheadi to be composed of four recognizable clades, each restricted to a geographic region within the Philippines. We consider K. bicolor, from peninsular Thailand, and K. pusilla, from Borneo, to be distinct from K. whiteheadi. Our data indicate the presence of two species within the Philippines currently lumped as K. hardwickii; further study of these is needed. A calibrated phylogeny suggests that Kerivoula began arriving in the Philippines about 10 MYA, with each of the four current lineages arriving independently.