crudely mapped as far as they were examined, and carefully inspected for bats and bat guano in 

 sections deemed safe to explore. Twelve workings were thus examined (Table 1 ), and data 

 loggers were left in six of these 



Bat detectors (ANABAT II, Titley Electronics, Ballina, Australia), mist nets, and/or harp 

 traps were deployed at workings where spoor was present or the mine working otherwise 

 appeared potentially suitable for bats Detector units (consisting of an ultrasound detector, 

 timer/tape-driver, and a voice-activated cassette tape recorder) were set before dusk facing 

 portals or aimed across shafts, and left in place overnight. Recorded calls were analyzed on an 

 IBM compatible PC using ANABAT II zero-crossings analysis interface module (ZCAIM) and 

 software. 



Assignment of vocalizations to a particular species of bat was achieved by matching 

 time-frequency structure of field recordings with a reference set of calls obtained from captured 

 individuals and published descriptions of vocalizations (e.g., Fenton et aJ. 1983, O'Farrell 1997) 

 However, bat species can show significant variation in call structure (Berts 1998, Barclay 1999), 

 and we did not actively track and record flying bats (O'Farrell et al. 1999) to maximize quality 

 and quantity of diagnostic sequences Furthermore, units recorded bats exiting roosts or flying 

 near potential roosts Roost-exit calls and calls in high clutter tend to be fragmentary, lacking 

 diagnostic features necessary for species identification (O'Farrell 1999). Therefore, all species- 

 level identifications based on recorded vocalizations, where made in this study, are considered 

 tentative. 



Myotis designations (as a group) were assigned to recordings with vocalizations of short 

 duration (< 3 msec) with a relatively linear, perpendicular call pattern. In some cases, Myotis 

 call sequences were assigned to M evotis if sweep pattern ranged from a maximum 90 kHz to a 

 minimum 35-40 kHz, otherwise all were classified Myotis species Calls with a bilinear (extreme 

 curvilinear) pattern were tentatively assigned to a non-Myotis species or classified as unknown 

 bat. Passes with call fragments were also designated unknown bat if no associated calls allowed 

 finer resolution Most bilinear call sequences were assigned to Eptesicus fuscus if a continuous 

 frequency tail ranged from 33-28 kHz This could result in confusion with Lasionyctens 

 noctivagans (Berts 1998), which has a similar call structure, but most of our recordings were 

 made at the mouths of mines where the latter species is unlikely to occur. 



Number of "passes" (defined here as a distinct vocalization with at least a 1 sec gap 

 between prior and following vocalizations) was recorded as a measure of relative activity at each 

 site. At five sites with bat activity, equipment malfunctioned prematurely Therefore, relative 

 activity as presented here is useful primarily as an index with variable degrees of error 



Bats were captured using 50-denier mist nets of various lengths (most often 6 and 9 

 meter) and set in a variety of arrays across portals, depending on site morphology Nets typically 

 were operated for at least three hours (usuallv until midnight or 01 00 MDT) Less frequently a 

 harp trap was set in the portal of an adit and left overnight Captured bats were identified with 

 aid of keys in van Zyll de Jong ( 1985) or Nagorsen and Brigham ( 1993) Individuals were 

 sexed, aged, measured (forearm, weight), reproductive status noted, then released. 



Where data are analyzed statistically, standard procedures and tests were followed as 

 described by Sokal and Rohlf ( 1981 ) G-tests were used to examine the null hypothesis of equal 

 proportions in frequency distributions, the null hypothesis of equal means in normally-distributed 

 data sets was examined using t-tests No particular probability level was assumed as 

 representing statistical significance, other than to consider a /-"-value of 05 or less to fall within 



