Photogrammetry (PH) is relatively cheap, easy to use, flexible and portable but its power and limitations have not been fully explored for studies of small animals. Here we assessed the accuracy of PH for the reconstruction of 3D digital models of bat skulls by evaluating its potential for evolutionary morphology studies at interspecific (19 species) level. Its reliability was assessed against the performance of micro CT scan (µCT) and laser scan techniques (LS). We used 3D geometric morphometrics and comparative methods to quantify the amount of size and shape variation due to the scanning technique and assess the strength of the biological signal in relation to both the technique error and phylogenetic uncertainty. We found only minor variation among techniques. Levels of random error (repeatability and procrustes variance) were similar in all techniques and no systematic error was observed (as evidenced from principal component analysis). Similar levels of phylogenetic signal, allometries and correlations with ecological variables (frequency of maximum energy and bite force) were detected among techniques. Phylogenetic uncertainty interacted with technique error but without affecting the biological conclusions driven by the evolutionary analyses. Our study confirms the accuracy of PH for the reconstruction of challenging specimens. These results encourage the use of PH as a reliable and highly accessible tool for the study of macro evolutionary processes of small mammals.

Morphological, functional, and behavioral adaptations of bats are among the most diverse within mammals. A strong association between bat skull morphology and feeding behavior has been suggested previously. However, morphological variation related to other drivers of adaptation, in particular echolocation, remains understudied. We assessed variation in skull morphology with respect to ecology (diet and emission type) and function (bite force, masticatory muscles and echolocation characteristics) using geometric morphometrics and comparative methods. Our study suggests that variation in skull shape of 10 bat families is the result of adaptations to broad dietary categories and sound emission types (oral or nasal). Skull shape correlates with echolocation parameters only in a subsample of insectivorous species, possibly because they (almost) entirely rely on this sensory system for locating and capturing prey. Insectivores emitting low frequencies are characterized by a ventrally tilted rostrum, a trait not associated with feeding parameters. This result questions the validity of a trade‐off between feeding and echolocation function. Our study advances understanding of the relationship between skull morphology and specific features of echolocation and suggests that evolutionary constraints due to echolocation may differ between different groups within the Chiroptera.

When selecting feeding, hiding, or resting areas, animals face multiple decisions with different fitness consequences. To maximize efficiency, individuals can either collect personal information, or use information gathered and transmitted by other individuals (social information). Within group living species, organisms often specialize in either generating social information or using information gathered by other groups members. That is the case of the Spix’s disk-winged bat, Thyroptera tricolor. This species uses contact calls during roost finding. Social groups are composed by a mix of vocal and non-vocal individuals and those vocal roles appear to be consistent over time. Moreover, their vocal behavior can predict roost finding in natural settings, suggesting that vocal individuals are capable of generating social information that can be used by other group members. To date, however, we do not know if when presented with social information (contact calls) during roost finding, vocal individuals will make more or less use of these cues, compared to non-vocal individuals. To answer this question, we broadcast contact calls from a roost inside a flight cage to test whether vocal individuals could find a potential roost faster than non-vocal individuals when they encounter sounds that signal the presence of a roost site. Our results suggest that non-vocal individuals select roost sites based primarily on social information, whereas vocal individuals do not rely heavily on social information when deciding where to roost. This study provides the first link between vocal behavior and the use of social information during the search for roosting resources in bats. Incorporating ideas of social roles, and how individuals decide when and where to move based on the use of social information, may shed some light on these and other outstanding questions about the social lives of bats.

Acoustic monitoring is an effective and scalable way to assess the health of important bioindicators like bats in the wild. However, the large amounts of resulting noisy data requires accurate tools for automatically determining the presence of different species of interest. Machine learning-based solutions offer the potential to reliably perform this task, but can require expertise in order to train and deploy. We propose BatDetect2, a novel deep learning-based pipeline for jointly detecting and classifying bat species from acoustic data. Distinct from existing deep learning-based acoustic methods, BatDetect2's outputs are interpretable as they directly indicate at what time and frequency a predicted echolocation call occurs. BatDetect2 also makes use of surrounding temporal information in order to improve its predictions, while still remaining computationally efficient at deployment time. We present experiments on five challenging datasets, from four distinct geographical regions (UK, Mexico, Australia, and Brazil). BatDetect2 results in a mean average precision of 0.88 for a dataset containing 17 bat species from the UK. This is significantly better than the 0.71 obtained by a traditional call parameter extraction baseline method. We show that the same pipeline, without any modifications, can be applied to acoustic data from different regions with different species compositions. The data annotation, model training, and evaluation tools proposed will enable practitioners to easily develop and deploy their own models. BatDetect2 lowers the barrier to entry preventing researchers from availing of effective deep learning bat acoustic classifiers.

Individuals within both moving and stationary groups arrange themselves in a predictable manner; for example, some individuals are consistently found at the front of the group or in the periphery and others in the center. Each position may be associated with various costs, such as greater exposure to predators, and benefits, such as preferential access to food. In social bats, we would expect a similar consistent arrangement for groups at roost-sites, which is where these mammals spend the largest portion of their lives. Here we study the relative position of individuals within a roost-site and establish if sex, age, and vocal behavior are associated with a given position. We focus on the highly cohesive and mobile social groups found in Spix's disc-winged bats (Thyroptera tricolor) given this species' use of a tubular roosting structure that forces individuals to be arranged linearly within its internal space. We obtained high scores for linearity measures, particularly for the top and bottom positions, indicating that bats position themselves in a predictable way despite constant roost-switching. We also found that sex and age were associated with the use of certain positions within the roost; for example, males and subadults tend to occupy the top part (near the roost's entrance) more often than expected by chance. Previous studies have shown that communally-roosting species often scramble to gain access to central positions, which are typically occupied by dominant individuals; thus, we speculate that our findings could also indicate some form of dominance hierarchy in our study species.