Do echolocation calls of wild colony-living Bechstein's bats (Myotis bechsteinii) provide individual-specific signatures?

Abstract

Social animals often use vocal communication signals that contain individual signatures. As bats emit echolocation calls several times per second to orient in space, these might seem ideal candidates for conveying the caller's individual identity as a free by-product. From a proximate perspective, however, coding of caller identity is hampered by the simple acoustic structure of echolocation signals, by their task-specific design and by propagation loss. We investigated the occurrence of individual signatures in echolocation calls in individually marked, free-living Bechstein's bats (Myotis bechsteinii) in a situation with defined social context in the field. The bats belonged to two different colonies, for both of which genetic data on relatedness structure was available. While our data clearly demonstrate situation specificity of call structure, the evidence for individual-specific signatures was relatively weak. We could not identify a robust and simple parameter that would convey the caller's identity despite the situation-specific call variability. Discriminant function analysis assigned calls to call sequences with good performance, but worsened drastically when tested with other sequences from the same bats. Therefore, we caution against concluding from a satisfactory discrimination performance with identical training and test sequences that individual bats can reliably be told apart by echolocation calls. At least the information contained in a single call sequence seems not to be sufficient for that purpose. Starting frequencies did give the best discrimination between individuals, and it was also this parameter that was correlated with genetic relatedness in one of our two study colonies. Echolocation calls could serve as an additional source of information for individual recognition in Bechstein's bats societies, while it is unlikely that a large number of individuals could be reliably identified in different situations based on echolocation alone.