Bats are a group of mammals well known for forming dynamic social groups. Studies of bat social structures are often based upon the frequency at which bats occupy the same roosts because observing bats directly is not always possible. However, it is not always clear how closely bats occupying the same roost associate with each other, obscuring whether associations result from social relationships or factors such as shared preferences for roosts. Our goal was to determine if bats cohabitating buildings were also found together inside roosts by using anti-collision technology for PIT tags, which enables simultaneous detection of multiple tags. We PIT-tagged 293 female little brown myotis (Myotis lucifugus) and installed antennas within two buildings used as maternity roosts in Yellowstone National Park. Antennas were positioned at roost entryways to generate cohabitation networks and along regions of attic ceilings in each building to generate intra-roost networks based on proximity of bats to each other. We found that intra-roost and cohabitation networks of buildings were significantly correlated, with the same bats tending to be linked in both networks, but that bats cohabitating the same building often roosted apart, leading to differing assessments of social structure. Cohabitation rates implied that bats associate with a greater number of their roost-mates than was supported by observations within the roost. This caused social networks built upon roost cohabitation rates to be denser, smaller in diameter, and contain nodes with higher average degree centrality. These results show that roost cohabitation does not reflect preference for roost-mates in little brown myotis, as is often inferred from similar studies, and that social network analyses based on cohabitation may provide misleading results.

Data were collected by scanning bats marked with uniquely identifiable PIT-tags using RFID antennas. Antennas were placed inside building roosts known to be used by female little brown myotis. For further details, see the referenced publication. 

The uploaded data at one-mode bat-by-bat matrices from each of the six networks constructed for the publication. As the publication describes, there are two roosts: one in Lamar Valley and one in Mammoth. There are two files for Lamar Valley because this roost was successfully studied in 2017 and 2018. Each file corresponds to one year of data. The Mammoth roost was only studied in 2018. Within each file there are two tabs. One contains the bat-by-bat matrix for data collected by antennas on the roost ceiling (intra-roost) and the other for data collected by antennas at roost entrayways (cohabitation). In each tab, rows and columns contain unique 4-character alphanumeric identifiers corresponding a bat. For the cohabitation matrices, the numbers in each cell correspond to the number of days a dyad (the combination of the bat identifiers at the intersection of a row and column) spent roosting togeher. For intra-roost matrices, the numbers in each cell correspond to hours detected at the same antenna.