Geographic variation in the echolocation calls of the endemic Cape horseshoe bat, Rhinolophus capensis (Chiroptera: Rhinolophidae)

Includes bibliographical references (leaves 63-84).Several intrinsic (body size) and extrinsic (foraging ecology and communication) factors are suggested to influence call frequency divergence in high duty-cycle bats. Investigating these factors within the framework of established hypotheses would contribute to understanding evolutionary changes leading to speciation in bats. Here, acoustic divergence between populations of the endemic Cape horseshoe bat, Rhinolophus capensis was investigated at both inter- and intraspecific levels. No previous study has investigated geographic variation in echolocation calls of R. capensis. Body size, wing morphology and skull parameters associated with diet and echolocation call production and reception, were compared between populations. Adult R. capensis were sampled at three sites: De Hoop situated in the centre of the species distribution in the Fynbos biome; Steenkampskraal and Table Farm were ecotone populations situated in the western and eastern limits of the distribution, respectively. Interspecific analysis revealed that the two ecotone populations deviated slightly from the allometric relationship between body size and peak frequency for the African clade. In fact, the expected inverse relationship between body size and peak frequency was not evident across populations. Ecotone populations had significantly larger mean body sizes than the population at De Hoop (10.28 ± 1.08 g; 84.60 ± 0.82 kHz). However, one population in the ecotone had the highest frequency (Table Farm: 13.88 ± 0.87 g; 85.84 ± 0.73 kHz) while the other had the lowest (Steenkampskraal: 13.15 ± 0.95 g; 80.66 ± 0.50 kHz). Several hypotheses were considered to explain the patterns of echolocation and morphological variation observed. The larger body size of the ecotone populations may be explained by James' Rule or it may be an adaptation to the intrinsic habitat heterogeneity of ecotones as it affords these bats a greater niche width and possibly larger home ranges to access spatially separated resources. On the other hand, neither climatic (humidity hypothesis), habitat (foraging habitat hypothesis) nor dietary differences (prey detection hypothesis between populations were responsible for the observed peak frequency differences between populations. Nasal chamber area was the best predictor of peak frequency and there was no relationship between the area of the nasal chamber and body size. Thus, selection may have acted directly on peak frequency altering skull parameters directly involved in echolocation independently of body size. Within each population, females were larger and used higher frequencies than males, which implies a potential social role of peak frequency for R. capensis. Observed differences in peak frequency may be because R. capensis interacts with separate rhinolophid species at either end of its distribution (Steenkampskraal: R. swinnyi; Table Farm: R. darlingi) in addition to R. clivasus, which results in the evolution of local dialects to facilitate intraspecific communication. These local dialects, possibly brought about by differences in local ambient noise characteristics (e.g. chorusing insects), could be maintained via cultural transmission. However, the role of gene flow for the evolution of these local dialects between populations cannot be discounted without adequate genetic analyses