Echolocating bats use and adapt ultrasound pulses that vary in several parameters, one of them being the pulse’s source level, which is a measure of the emitted sound amplitude. This is a vital ecological parameter as it directly impacts the maximum distance over which bats can perceive targets in their environment, most importantly their prey. Different habitats present different sensing challenges for echolocation systems, and the quality and content of information derived from echolocation pulses reflect these environmental challenges. As such, echolocation pulses within or between species may vary from one habitat to the next due to variable selection pressure, resulting in local adaptation. Habitat is, therefore, a key component in shaping the evolution of echolocation. The Acoustic Adaptation Hypothesis (AAH) proposes that acoustic properties of the environment influence sound propagation and therefore the evolution of echolocation pulses. Here, we tested the AAH using multiple microphone arrays to measure the source levels of echolocation pulses of 14 bat species in bat assemblages across sites in six biomes in South Africa. Contrary to the AAH, our results revealed that bats in the same assemblage used different echolocation pulse source levels, frequencies, and duration resulting in different detection distances, which differ among bat assemblages occupying different sites. Furthermore, detection distance was species-specific and remained similar within species between assemblages; suggesting that species is a better predictor of detection distances compared to habitat as indicated by Miniopterus natalensis across all seven sites. KEYWORDS: adaptation, bat assemblages, detection distances, microphone arrays, selection pressure, source levels

The relative contributions of adaptation and drift to morphological diversification of the crania of echolocating mammals was investigated using two horseshoe bat species, Rhinolophus simulator and R. cf. simulator as test cases. We used 3D geometric morphometrics to compare the shapes of skulls of the two lineages collected at various localities in southern Africa. Shape variation was predominantly attributed to selective forces; the between population variance (B) was not proportional to the within population variance (W). Modularity was evident in the crania of R. simulator but absent in the crania of R. cf. simulator and the mandibles of both species. The skulls of the two lineages thus appeared to be under different selection pressures, despite the overlap in their distributions. Selection acted mainly on the nasal dome region of R. cf. simulator whereas selection acted more on the cranium and mandibles than on the nasal domes of R. simulator. Probably the relatively higher echolocation frequencies used by R. cf. simulator, the shape of the nasal dome, which acts as a frequency dependent acoustic horn, is more crucial than in R. simulator, allowing maximization of the intensity of the emitted calls and resulting in comparable detection distances. In contrast, selection pressure is probably more pronounced on the mandibles and cranium of R. simulator to compensate for the loss in bite force because of its elongated rostrum. The predominance of selection probably reflects the stringent association between environment and the optimal functioning of phenotypic characters associated with echolocation and feeding in bats.