2.3. Bat sampling

Bats were captured using ground‐level mist nets during the dry season, between July and November of 2011 and 2012, and the wet season, between February to June of 2012 and 2013. Each interior site was surveyed eight times, four times in each season. The number of visits to edge and matrix sites ranged from 3 to 6 in the wet season and 2 to 3 in the dry season. For each survey, 14 mist nets (12 × 2.5 m, 16 mm mesh, ECOTONE, Poland) were used in continuous forest and fragment interiors, and seven mist nets at the edge and adjacent matrix sites. Nets were left open during 6 hr from dusk to midnight and were visited at intervals of ~20 min. The same site was never surveyed during two consecutive nights to avoid net shyness‐related capture bias (Marques et al., 2013). Adult bats (excluding pregnant females) were marked with numbered ball‐chain necklaces (Pteronotus parnellii and frugivores) or transponders (gleaning animalivores). Species identification followed Gardner (2008) and Lim and Engstrom (2001), and taxonomy follows Gardner (2008). The analyses were limited to phyllostomids and P. parnellii due to under‐representation of other families and species with this type of sampling method (Kalko, 1998).