RESULTS AND DISCUSSION 

 Bat Detector Surveys 



Species Richness— A total of 54 sites in 1994 and 134 sites in 1995 was checked for nocturnal 

 bat activity using ultrasound bat detectors: 9 sites in July, 63 in August, and 1 16 in September. 

 Number of sites surveyed on each District, with years combined, was Cabinet (34), Fisher River 

 (28), Fortine (32), Libby (28), Rexford (2 1 ), and Three Rivers (45) (see Appendix 2). All sites 

 were sampled with ANABAT II ultrasonic bat detectors. Bats were detected at 133 (70.7%) of 

 the sites: 40 (74.1%) in 1994, 93 (69.4%) in 1995. The proportion of sites where bats were 

 detected did not differ (G = 0.412, df= \,P> 0.5) between years . 



Five species of vespertilionid bat (common names usually follow Jones et al. 1986) were 

 identified using ultrasound bat detectors each year during the field surveys (see Appendix 3): 

 long-eared myotis {Myotis evotis - 28 sites), silver-haired bat {Lasionycteris noctivagans - 45 

 sites), big brown bat {Eptesicus fuscus -57 sites), hoary bat (Lasiurus cinereus - 13 sites), 

 Townsend's big-eared bat {Corynorhinus townsendii - 36 sites). In addition, Myotis sp. was 

 detected at 107 sites. 



Number of bat species detected at a site ranged fi-om one to five (assuming Myotis sp. 

 represented a single species other than M. evotis at each site where detected). One species was 

 detected at 34 sites, two species at 35 sites, three species at 36 sites, four species at 17 sites, and 

 five species at 1 1 sites; a mean of 2.5 ±1.2 species/site (= species richness) was detected at sites 

 where bats were present during the two years. Species richness was 3.1 ± 1.3 in 1994 and 2.3 ± 

 1.1 in 1995; fi-equency distributions of species richness differed significantly between years (G = 

 13.616, df= 4, P< 0.01). 



Several explanations for the difference in species richness between years are possible. First, 

 distinguishing bat calls may have been more conservative in 1995, as experience with their 

 interpretation increased. We have no way to check this, so we will not discuss this possibility 

 further except to note that observer bias is always present to some degree. Second, bat activity 

 may have been depressed in 1 995 due to some sampling artifact, such as inclement or colder 

 weather on nights sampled. Neither wind nor sky conditions at the time bat detectors were 

 placed differed significantly between years (G tests, P > 0.1), but ambient temperature on 

 evenings sampled was significantly warmer in 1994 (mean = 20.0 + 3.2 °C in 1994, 16.2 ± 3.0 °C 

 in 1995; t = 3.887, df = 40, P < 0.001); bat and night-flying insect activity may be correlated 

 with nocturnal ambient temperature in northwestern Montana. Third, the array of habitats 

 surveyed may have occurred in different proportions each year; greater species richness in 1994 

 may have been the result of a greater proportion of samples in habitats "attractive" to a greater 

 number and variety of bats. The proportion of gross habitat categories sampled (old- 

 growth/mature, "disturbed", all other habitats including riparian) did not differ significantly 

 between years (G = 4.566, df = 2, P > 0.1), but pooling habitats at this level could mask finer- 

 scale habitat selection by bats that is biologically significant. In the absence of evidence to the 

 contrary, we assume that the difference between years in species richness was a real property of 

 the environment. 



