Augmentation of Two Identification Methods for Bats

Abstract

All marking methods for identifying bats (order Chiropteran) have practical limitations, with no one method being superior to others. To address these limitations, we proposed and tested the use of two prospective identification methods p-Chip microtransponder tags and the use of collagen-elastin (CE) bundle patterns as a biomarker in a captive colony of big brown bats (Eptesicus fuscus). For p-Chips, we assessed (1) animal handling time, (2) scan time, (3) number of wand flashes, (4) p-Chip visibility, (5) readability, and (6) the bat’s overall condition for two locations: all bats had p-Chips implanted in the wing (n = 30) and some of these bats also had p-Chips implanted in their leg (n = 13). For both locations, average scan times increased over time whereas the number of wand flashes decreased, suggesting p-Chip recording efficacy improves with user experience. The visibility and readability of p-Chips was consistently better for tags injected in the wing compared the leg, emphasizing the wing as the preferred implantation site. A second proposed identification method extends upon the use of manual, visual inspection (Amelon et al. 2017) to examine whether pattern-recognition software can accurately detect and identify individual bats using the pattern of collagen and elastin bundles in the wing. We tested the effectiveness of HotSpotter© to identify adult (n = 24 bats; n = 192 photos) and juvenile (n = 34 pups; n = 136 photos) E. fuscus by comparing photos of the wing membrane illuminated by ultraviolet light. We then assessed similarity scores between adults and juveniles separately and quantified the occurrence of correct and incorrect matches. For images of adult bats, 60% of comparisons resulted in a correctly matched top-ranked image (i.e. an image of the same bat was most similar), whereas 27% of comparisons had a correct top-ranked image for wing membrane photos of juvenile bats. The success rate of obtaining a correct match could be increased by including a larger subset of top-ranked images when selecting possible correct matches. Altogether, these results suggest that p-Chip tags and potentially the use of HotSpotter pattern recognition software are suitable methods for identifying captive E. fuscus and may be viable for use in the field and in other bat species.