Temperature-sensitive dataloggers were programmed to read skin temperature (Tsk) every 30 min and were attached to 504 bats over the course of three winters at six different hibernacula using standard methods [22] . Temperature readings could not be collected more frequently due to constraints on datalogger memory and the need to record continuous data for up to five months. To maximize recapture rates, bats with loggers were recaptured in March of each year, several weeks prior to the ‘normal’ time of emergence from hibernation. Loggers weighted about 1.1 g and were either purchased commercially (iBBat or WeeTagLites, AlphaMach, Inc., British Columbia, Canada) or were constructed by the authors (DMR and GGT). Appendix S1 describes and illustrates the methods for making these dataloggers from Thermochron DS1922L iButtons (Maxim Integrated Products, Inc., California, USA), modified from the techniques of Lovegrove [23] . Table 1 provides a summary of loggers deployed, retrieved, and downloaded successfully, by site, year, and sex.
Study sites were widely distributed and located in Vermont, West Virginia, Pennsylvania, and the Upper Peninsula of Michigan (Fig. 1). Among loggers retrieved, success rates varied. WeeTagLites failed at a rate of up to 90% whereas loggers constructed by the authors failed about 20% of the time. Overall 111 of 190 loggers retrieved yielded usable data, an average of 58.4%. We expected to recover less than half the loggers placed in the field and expected datalogger failure as well, which is why so many loggers were deployed. Of the 190 bats from which loggers were retrieved, 17 were found dead (four of which were in suitable post-mortem condition to perform histology analysis). For the 173 live bats recaptured in the spring, loggers were removed, and the animal was either released (N = 126) or euthanized for measurement of immune function and other physiological parameters for a separate study (N = 25) or for histology analysis (N = 22), as described below.
Free full text: Click here