Ears evolved in many nocturnal insects, including some moths, to detect bat echolocation calls and evade capture [1, 2]. Although there is evidence that some bats emit echolocation calls that are inconspicuous to eared moths, it is difficult to determine whether this was an adaptation to moth hearing or originally evolved for a different purpose [2, 3]. Aerial-hawking bats generally emit high-amplitude echolocation calls to maximize detection range [4, 5]. Here we present the first example of an echolocation counterstrategy to overcome prey hearing at the cost of reduced detection distance. We combined comparative bat flight-path tracking and moth neurophysiology with fecal DNA analysis to show that the barbastelle, Barbastella barbastellus, emits calls that are 10 to 100 times lower in amplitude than those of other aerial-hawking bats, remains undetected by moths until close, and captures mainly eared moths. Model calculations demonstrate that only bats emitting such low-amplitude calls hear moth echoes before their calls are conspicuous to moths. This stealth echolocation allows the barbastelle to exploit food resources that are difficult to catch for other aerial-hawking bats emitting calls of greater amplitude.

Echolocation in bats and high-frequency hearing in their insect prey make bats and insects an ideal system for studying the sensory ecology and neuroethology of predator-prey interactions. Here, we review the evolutionary history of bats and eared insects, focusing on the insect order Lepidoptera, and consider the evidence for antipredator adaptations and predator counter-adaptations. Ears evolved in a remarkable number of body locations across insects, with the original selection pressure for ears differing between groups. Although cause and effect are difficult to determine, correlations between hearing and life history strategies in moths provide evidence for how these two variables influence each other. We consider life history variables such as size, sex, circadian and seasonal activity patterns, geographic range and the composition of sympatric bat communities. We also review hypotheses on the neural basis for antipredator behaviours (such as evasive flight and sound production) in moths. It is assumed that these prey adaptations would select for counter-adaptations in predatory bats. We suggest two levels of support for classifying bat traits as counter-adaptations: traits that allow bats to eat more eared prey than expected based on their availability in the environment provide a low level of support for counter-adaptations, whereas traits that have no other plausible explanation for their origination and maintenance than capturing defended prey constitute a high level of support. Specific predator counter-adaptations include calling at frequencies outside the sensitivity range of most eared prey, changing the pattern and frequency of echolocation calls during prey pursuit, and quiet, or 'stealth', echolocation.

Evidence suggests that behavioural defences, such as habitat selection and grooming behaviour, have evolved in animals in response to the costs associated with ectoparasites. Bat fly and mite densities were compared among wild-caught bats in Belize with different roosting preferences (cavity, foliage, or both), and grooming behaviour was analysed for bat species with high and low ectoparasite density. Ectoparasites of bats were removed using forceps, and bat grooming behaviour was recorded with a camcorder. Because bat flies pupate on the surface of host roosts, bats that use cavity roosts (a sheltered environment for the pupae) were predicted to have higher densities of bat flies than those that use foliage (exposed environment). Cavity-roosting species generally had higher densities of bat flies and mites, although the relationship was more evident for bat flies. The grooming behaviour of bats was predicted to differ among species with high or low ectoparasite densities. Although there was no difference in the frequency of grooming behaviours for individuals with and without bat flies, there were differences in grooming behaviour at the species level. Bat species with high ectoparasite densities scratched more than those with low ectoparasite densities. These results suggest that ectoparasite densities and grooming behaviour are related to roosting preferences in bats.

New communication signals can evolve by sensory exploitation if signaling taps into preexisting sensory biases in receivers [1, 2]. For mate attraction, signals are typically similar to attractive environmental cues like food [3-6], which amplifies their attractiveness to mates, as opposed to aversive stimuli like predator cues. Female field crickets approach the low-frequency calling song of males, whereas they avoid high-frequency sounds like predatory bat calls [7]. In one group of crickets (Eneopterinae: Lebinthini), however, males produce exceptionally high-frequency calling songs in the range of bat calls [8], a surprising signal in the context of mate attraction. We found that female lebinthines, instead of approaching singing males, produce vibrational responses after male calls, and males track the source of vibrations to find females. We also demonstrate that field cricket species closely related to the Lebinthini show an acoustic startle response to high-frequency sounds that generates substrate vibrations similar to those produced by female lebinthine crickets. Therefore, the startle response is the most likely evolutionary origin of the female lebinthine vibrational signal. In field crickets, the brain receives activity from two auditory interneurons; AN1 tuned to male calling song controls positive phonotaxis, and AN2 tuned to high-frequency bat calls triggers negative phonotaxis [9, 10]. In lebinthine crickets, however, we found that auditory ascending neurons are only tuned to high-frequency sounds, and their tuning matches the thresholds for female vibrational signals. Our results demonstrate how sensory exploitation of anti-predator behavior can evolve into a communication system that benefits both senders and receivers.

Ectoparasite host specificity can be influenced by factors such as the degree of host isolation and ectoparasite mobility. Host-site specificity can result from factors such as proximity to mates, competition, and host grooming behaviour. Ectoparasitic bat flies on bats from the Lamanai area of Belize were collected from hosts captured in mist nets to determine host specificity and host-site specificity. Bat grooming behaviour was also recorded and quantified. From 455 bats (25 species in five families), 773 bat flies (32 species in two families) were collected. Of 32 bat fly species, 25 were only found on 1 bat species, 6 were found on 2 species of the same genus, and 1 was found on 2 species of different genera (the latter appearing to be an accidental association). Specificity of the bat flies tended to follow the taxonomy of the bat hosts, not the ecological isolation of the host species, since bat species that often roost in polyspecific groups did not share bat fly species. Mobility of the bat flies was not related to host specificity. Host-site specificity of bat flies occurred for either fur or membrane on the host, and long hind legs and ctenidia appear to be morphological adaptations for living in fur. Bat grooming behaviour was consistent with the assumptions of a simulation model, which suggested that host grooming could be responsible for host-site segregation of bat flies.

The Brazilian free-tailed bat, Tadarida brasiliensis (Saint-Hilaire, 1824), uses calls that represent a broad continuum of design variation which is dependent upon habitat and situation, and exhibits characteristic changes in call design as bats close in on airborne targets. Here we demonstrate the influence of conspecifics on call design. We found that the peak frequency used in calls varies more as the number of bats flying in the same space increases (measured from single bats and pairs of bats). We investigated this phenomenon through comparing call-parameter differences found between two bats recorded flying together (actual pairs) with call-parameter differences between two bats each recorded flying alone at different locations that were randomly assigned to one another (virtual pairs). We found that actual pairs of bats used calls which differed in peak frequency more so than did virtual pairs. This result is particularly striking given that these frequency differences were greater between bats in the same space than between bats in two different habitats. We argue that these differences indicate that this species is practicing jamming avoidance, air traffic control, or both.

Many predators and parasites eavesdrop on the communication signals of their prey. Eavesdropping is typically studied as dyadic predator-prey species interactions; yet in nature, most predators target multiple prey species and most prey must evade multiple predator species. The impact of predator communities on prey signal evolution is not well understood. Predators could converge in their preferences for conspicuous signal properties, generating competition among predators and natural selection on particular prey signal features. Alternatively, predator species could vary in their preferences for prey signal properties, resulting in sensory-based niche partitioning of prey resources. In the Neotropics, many substrate-gleaning bats use the mate-attraction songs of male katydids to locate them as prey. We studied mechanisms of niche partitioning in four substrate-gleaning bat species and found they are similar in morphology, echolocation signal design and prey-handling ability, but each species preferred different acoustic features of male song in 12 sympatric katydid species. This divergence in predator preference probably contributes to the coexistence of many substrate-gleaning bat species in the Neotropics, and the substantial diversity in the mate-attraction signals of katydids. Our results provide insight into how multiple eavesdropping predator species might influence prey signal evolution through sensory-based niche partitioning.

The short-tailed fruit bat, Carollia perspicillata, lives in groups in tree hollows and caves. To investigate whether these roosts might serve as information centres, we tested whether individuals' preferences for novel foods could be enhanced through social learning at the roost. We also determined whether socially learned preferences for novel foods were reversed through interaction with other roost mates by simulating changes in available food resources such as those associated with variations in timing of fruit production in different plant species. Bats exhibited socially induced preferences that were readily reversible. We suggest that for frugivorous bats, roosts can serve as centres for information exchange about novel and familiar, ephemeral foods without requiring conspecific recruitment to these resources.