Effects of interindividual variation in echolocation calls on identification of big brown and silver-haired bats

Identifying bat species from their calls is a common technique in studies of habitat use. I tested the hypotheses that significant interindividual intraspecific variation exists in the echolocation calls of bats as detected by frequency-divide ultrasonic detectors, and that such variation could greatly affect die accuracy of species identification. I restricted analysis to 2 species that have similar but reportedly distinguishable calls. There was significant intraspecific variation in 6 variables for each species, and frequency distributions of the 2 species overlapped greatly for all 6 variables; hence, univariate analysis was ineffective at differentiating between species. The best of 4 experienced human observers correctly identified the species for only 70% of 47 call sequences compared to >95% correct classification by discriminant function analysis. Alterations in current methodology can improve accuracy of identification and thus the value of habitat studies and manage ment decisions that rely on such identification.