Northeastern Naturalist 72 S.J. Wolbert, A.S. Zellner, and H.P. Whidden 22001144 NORTHEASTERN NATURALIST V2o1l.( 12)1:,7 N2–o8. 51 Bat Activity, Insect Biomass, and Temperature Along an Elevational Gradient Stacy J. Wolbert1, Andrew S. Zellner2, and Howard P. Whidden3,* Abstract - Recent studies have documented high levels of bat fatalities at wind-energy facilities built on forested ridgetops in the eastern United States. To better understand the influence of temperature and elevation on bat activity levels, and the possible relationship of these factors to bat fatalities at wind facilities, we sampled bat activity and insect abundance along an elevational gradient at 3 study areas in northeastern Pennsylvania. Bat activity was sampled with an AR125 acoustic detector, and insect abundance was sampled with a blacklight trap. We developed a negative binomial regression model to assess the relationship between bat activity and temperature, elevation, and insect abundance. We also assessed a hypothesized association between the occurrence of temperature inversions and increased bat activity at higher elevations. We found a significant positive association between bat activity and temperature (P < 0.001), with the effect of temperature being greater at higher elevations (P = 0.021). Contrary to predictions, there was a significant negative relationship between bat activity and insect biomass (P < 0.001), and the association between bat activity and the occurrence of temperature inversions was not significant (P = 0.1). Although we did find significantly greater bat activity at higher temperatures, and an interaction between temperature and elevation, our results do not support temperature inversions as a factor in bat fatalities at wind-energy facilities on forested ridgetops in the eastern US. Introduction Although commercial wind-energy facilities have been around for many years, their role in causing bat mortality was widely recognized only in 2003, when large numbers of bat kills were documented at the Mountaineer Wind Energy Center in Tucker County, WV (Kerns and Kerlinger 2004). Subsequent studies have found that bat mortality occurs regularly at wind facilities throughout the eastern United States (Arnett et al. 2008, Kunz et al. 2007a), although the factors that influence bat mortality rates are still not well understood (Cryan and Barclay 2009, Johnson et al. 2003, Kunz et al. 2007b). Mortality surveys at wind facilities consistently find high levels of mortality in migratory bats such as Lasiurus cinereus Beauvois (Hoary Bat), Lasiurus borealis Müller (Eastern Red Bat), and Lasionycteris noctivagans LeConte (Silver-haired Bat), whereas hibernating bats generally have much lower mortality rates (Arnett et al. 2008; Johnson et al. 2003, 2004). Somewhat surprisingly, high levels of mortality at wind facilities are also reported for Perimyotis subflavus F. Cuvier (Tri-colored Bat) (Arnett et al. 2008). This species hibernates during the winter and has traditionally been con- 12001 Elmerton Avenue, Harrisburg, PA 17110. 2163 Tech Pointe Drive, Fitzgerald, GA 31750. 3East Stroudsburg University, East Stroudsburg, PA, 18301. *Corresponding author - hwhidden@esu.edu. Manuscript Editor: Peter Paton Northeastern Naturalist Vol. 21, No. 1 S.J. Wolbert, A.S. Zellner, and H.P. Whidden 2014 73 sidered a short-distance regional migrant, although recent stable-isotope evidence suggests that at least some individuals undergo a significant southward migration in the fall (Fraser et al. 2012). All bats native to Pennsylvania are insectivorous, although species differ in their foraging styles and preferred insect prey (Whitaker 2004, Whitaker and Hamilton 1998). Despite these differences, most species appear to be opportunistic feeders (Carter et al. 2004, Kurta and Whitaker 1998). Because of this opportunism, bats are predicted to shift among patches of insects in response to geographic and temporal variation in prey availability (Arbuthnott and Brigham 2007, Lacki et al. 1995, Moosman et al. 2012). Temperature also affects bat-activity patterns, and numerous studies have documented a positive relationship between bat activity and temperature, with a minimum temperature threshold below which activity rapidly declines (Agosta et al. 2005, Reynolds 2006, Rydell et al. 1996). Thermal inversions may also have an effect on bat activity patterns. These inversions create cool, foggy conditions in valleys with warmer air masses rising to ridgetops (Bates and Jackson 1984, Kunz et al. 2007b). Because insect activity also increases with temperature (Mellanby 1939, Raimondo et al. 2004, Taylor 1963), both bats and insects are hypothesized to increase their activity along ridgetops in response to temperature inversions (Kunz et al. 2007b). Wind-energy facilities in the eastern United States are commonly built along forested ridgetops (Arnett et al. 2008), and the hypothesized shift in bat activity to ridgetops during temperature inversions may help explain the high fatality rates for bats at wind-energy facilities in the eastern US (Kunz et al. 2007b). To better understand the role of temperature and elevation in determining batactivity levels, and assess the possible association of these factors with bat mortality at wind facilities, we sampled bat activity, insect abundance, and temperature along an elevational gradient at three study areas in northeastern Pennsylvania. The objectives of this study were to: (1) assess the relationship between ambient temperature and bat activity along an elevational gradient, (2) investigate associations between levels of bat activity and the availability of insect prey, and (3) assess the impact of temperature inversions on insect abundance and bat activity along ridgetops. We hypothesized that temperature inversions would increase insect availability at highelevation sites and in turn lead to increased bat activity. Study Area and Methods Study areas We sampled bat activity and insect biomass along an elevational gradient at three study areas in northeastern Pennsylvania (Fig. 1): Crystal Lake in Luzerne County (41.19°N, 75.88°W), Hickory Run State Park in Carbon County (41.02°N, 75.69°W), and Hypsie Gap in Monroe County (41.01°N, 75.42°W). We chose these study areas because they each had sampling sites at three elevations (335 m, 457 m, and 579 m) that were easily accessible from a public road. The Crystal Lake study area was in Luzerne County approximately 7 km south-southeast of the city of Wilkes-Barre. The area was forested, with the lower-elevation sampling site dominated by Quercus prinus L. (Chestnut Oak), Northeastern Naturalist 74 S.J. Wolbert, A.S. Zellner, and H.P. Whidden 2014 Vol. 21, No. 1 Quercus rubra L. (Northern Red Oak), and Acer rubrum L. (Red Maple), the middle-elevation sampling site dominated by Rhus typhina L. (Staghorn Sumac) and Fagus grandifolia Ehrh. (American Beech) saplings, and the upper-elevation sampling site dominated by Betula alleghaniensis Britton (Yellow Birch) and Red Maple trees. The Hickory Run study area was in Hickory Run State Park in Carbon County, and the lower-elevation sampling site was immediately adjacent to Hickory Run, a stream with continuous year-round flow. Kalmia latifolia L. (Mountain Laurel) was abundant at all elevations, with Red Oak and American Beech common at the lower- and middle-elevation sampling sites, and Yellow Birch trees and Pinus strobus L. (White Pine) and Tsuga canadensis (L.) Carrière (Eastern Hemlock) saplings common at the upper-elevation sampling site. The Hypsie Gap study area was in Monroe County approximately 10 km west-southwest of Tannersville. The lower-elevation sampling site there was dominated by Rhododendron maximum L. (Rosebay Rhododendron), with Chestnut Oak trees and Red Maple, Red Oak, and Yellow Birch saplings. At the middle-elevation sampling site, the most abundant trees were Yellow Birch and Chestnut Oak, whereas the upper-elevation sampling site was dominated by Quercus ilicifolia Wangenh. (Bear Oak), with Sassafras albidum (Nutt.) Nees (Sassafras) trees and Red Maple and Eastern Hemlock saplings. Figure 1. Location of three study areas in northeastern Pennsylvania where bat activity and insect abundance were monitored during 2006. Northeastern Naturalist Vol. 21, No. 1 S.J. Wolbert, A.S. Zellner, and H.P. Whidden 2014 75 Methods We conducted fieldwork at each study area one night per week April–November 2006, with the three elevations at a study area sampled simultaneously. Using the roll of a 6-sided die, we randomized the order in which the three study areas were sampled in a given week. We began sampling at official sunset and continued for five hours, with sunset times obtained from the United States Naval Observatory (USNO), Astronomical Applications Department (http://www.usno.navy. mil). We used AR125 acoustic detectors (Binary Acoustic Technology, Tucson, AZ) to provide a measure of bat activity and insect light traps (model #2857A, BioQuip, Rancho Dominguez, CA) to sample insect populations. We placed the AR125 acoustic detectors oriented vertically 1.2 m off the ground to passively record bat activity. The program SPECT’R (Binary Acoustic Technology, Tucson, AZ) recorded echolocation calls as .wav files on a Gateway M460 laptop computer (Gateway, Inc., Irvine, CA). We set SPECT’R with the following parameters: (1) files were recorded as snapshots using the Autosnap option, (2) the Autosnap trigger threshold was 15.0 dB, (3) the Autosnap trigger range was 18–120 kHz, (4) the Autosnap duration was 1.0 second, and (5) the time expansion factor was 10. We then sorted the recorded files into 1-hour sampling intervals, which corresponded with insect-sampling periods. We analyzed recorded echolocation calls manually using SonoBat 2.5.3 (SonoBat, Arcata, CA) to count bat passes. We defined a bat pass as ≥1 echolocation pulse recorded within a 1.0 second time frame (Hayes 1997, Weller and Zabel 2002). We used the number of bat passes recorded per hour as an index of bat activity (Gannon et al. 2003, Hayes 1997, Sherwin et al. 2000). We did not identify bat passes to species. We sampled insects as per Lacki et al. (1995), with light traps placed on the ground 50–75 m from the AR125 acoustic detectors to avoid recording bats that were attracted to insects near the light traps. We emptied the light traps every hour and stored the hourly insect samples separately in plastic bags. Each hour, at intervals corresponding to the bat-sampling intervals, we used a rubber band to attach a new plastic bag to the funnel of each insect light trap. We used cotton balls soaked with ethyl acetate as the killing agents and placed one in each plastic collection bag during sampling. We froze the insect samples and later counted the insects and identified them to order using keys in Triplehorn and Johnson (2005). We determined insect biomass for each sample by drying at approximately 55 °C for 5 days (Barclay 1991). We grouped insects by order and weighed each group to the nearest 0.001 g on an electronic balance. We mounted Hobo H8 Pro Series data loggers (Onset Computer Corp., Pocasset, MA) on trees at the three elevations for each study area and set them to record temperature every 10 minutes. We averaged the 10-minute temperature readings to obtain mean hourly and nightly temperatures for each elevation. Statistical analysis Prior to analysis, we summed hourly data on bat passes, insect biomass, and insect numbers to obtain nightly values for each parameter. We also added a binary Northeastern Naturalist 76 S.J. Wolbert, A.S. Zellner, and H.P. Whidden 2014 Vol. 21, No. 1 variable for season, with two values, pre-volant (<15 July) and volant (≥15 July), to assess an expected late-summer increase in bat activity when juvenile bats began to fly independently. We assessed multicollinearity of variables by performing correlation analyses (Kendall’s τ) and determining generalized variance inflation factors (GVIFs) for our variables. Because several predictor variables were categorical with multiple degrees of freedom, we relied on GVIF1/(2*DF) to assess possible effects of collinearity (Fox and Monette 1992). Initial data exploration indicated that the distribution of bat passes was positively skewed and the variance was much greater than the mean. We therefore used generalized linear model procedures (glm and glm.nb) in R (version 3.0.1, R Core Team 2013) to compare the fit of log-linked Poisson, quasiPoisson, and negative binomial models to our data (Crawley 2013). The number of nightly bat passes was the response variable, and study area, elevation, season, temperature, insect biomass, and insect numbers were the predictor variables. We anticipated that there would be interactions between study area and elevation and between elevation and temperature, and therefore we included these interactions (study area:elevation and elevation:temperature) as additional predictor variables. The different regression models were compared using visual evaluation of quantile-quantile (QQ) plots, comparison of residual deviance vs. residual degrees of freedom, and assessment of Akaike’s information criterion (AIC) (Crawley 2013). After determining the best-fitting distributional model for the complete set of predictor variables, we evaluated hierarchical candidate models using the functions drop1 and stepAIC. To assess the effects of a temperature inversion on bat activity and insect biomass, we developed a separate negative binomial model to compare hourly trends on nights with and without a temperature inversion. We ran this model for the high-elevation (579 m) sites showing temperature inversions, with bat passes and insect biomass as separate response variables and hour, inversion (yes/no), and the interaction between hour and inversion (hour:inversion) as predictor variables. We then used a chi-square test to evaluate the significance of the interaction effects between hour and the presence/absence of a temperature inversion. Results We attempted sampling on 90 nights for 5 hours a night, resulting in a potential sampling effort of 1350 total hours. However, extreme weather conditions and equipment failure caused a loss of 31 hours at Crystal Lake (93% success), 35 hours at Hickory Run (92% success), and 47 hours at Hypsie Gap (90% success) for an overall loss of 113 hours (92% overall success rate). The final total was 1237 hours of acoustic monitoring and complete data (bat activity, insect biomass, insect numbers, and temperature) for 238 nightly elevational samples (Table 1). In the 1237 hours of acoustic monitoring, we recorded 67,575 bat passes (Table 1). Numbers of bat passes varied with both study area and elevation, and more than 70% of all bat passes were recorded at the lower-elevation (335 m) sampling site at Hickory Run (Fig. 2). Bat activity began to increase with ambient temperatures above 10 °C, with the highest levels of activity recorded through about 18 Northeastern Naturalist Vol. 21, No. 1 S.J. Wolbert, A.S. Zellner, and H.P. Whidden 2014 77 °C, although a large proportion of hourly samples at all temperatures showed little or no bat activity (Fig. 3a). We captured 70,166 insects in the light traps, with representatives of 10 orders of potential prey and a total biomass of 345.28 g. The most commonly captured orders by biomass were: Lepidoptera (220.46 g), Coleoptera (99.37 g), Trichoptera (9.96 g), Diptera (7.98 g), Hemiptera (4.42 g), Hymenoptera (2.58 g), Ephemeroptera (0.44 g), and Neuroptera (0.07 g). Insect biomass began to increase at about 8 °C, with high levels of biomass recorded through the maximum recorded temperatures of approximately 28 °C (Fig. 3b). A large proportion of hourly samples below 20 °C had little or no insect biomass, but all samples above 22 °C contained at least 0.18 g of insect biomass. Correlation analyses found low to moderate correlations (τ < 0.6) between all pairs of predictor variables except between insect numbers and temperature (τ = 0.617). All GVIF1/(2*DF) values were <4.0, with a mean of 2.03, suggesting only moderate inflation of standard errors due to collinearity of variables (Field et al. 2012). The regression analyses used the 238 nightly samples with complete data (bat passes, insect biomass, insect numbers, and temperature). The negative binomial model was a much better fit for our data than either Poisson or quasiPoisson models, as evidenced by the fit of the QQ plot, correspondence between residual deviance and residual degrees of freedom, and lower AIC scores (Poisson = 64,750; negative binomial = 2552). The drop1 function revealed that the number of insects was not a significant predictor of bat activity (P = 0.586), but all the other predictor variables were significant (Table 2). The stepAIC function led to a negative binomial model that included all initial predictor variables except insect numbers (AIC = 2548). The final regression model indicated significant variation in bat activity between the three study areas, with Hickory Run having the most bat activity, followed by Crystal Lake, and then Hypsie Gap. There was no significant difference in bat activity between lower- and middle-elevation sampling sites, but bat activity was significantly less at upper-elevation sites (P = 0.006). The effect of elevation varied among study areas, with the Hypsie Gap upper-elevation sampling site associated with increased bat activity and the Hickory Run middle- and upper-elevation sampling sites associated with decreased bat activity. Activity was also significantly less in the volant period compared with the pre-volant period. Temperature was a highly significant predictor of bat activity (P < 0.001), and the effect of temperature Table 1. Total number of bat passes recorded from 17 April–8 November 2006 at three elevations at each of three study areas in northeastern Pennsylvania. Total number of nights/hours of acoustic monitoring is shown in parentheses. Elevation Crystal Lake Hickory Run Hypsie Gap Total Lower (335 m) 8246 (28/140) 47,750 (27/135) 538 (27/135) 56,534 (82/410) Middle (457 m) 2428 (27/139) 2252 (26/140) 659 (28/133) 5339 (81/412) Upper (579 m) 3212 (28/140) 1815 (27/140) 675 (20/135) 5702 (75/415) Total 13,886 (83/419) 51,817 (80/415) 1872 (75/403) 67,575 (238/1237) Northeastern Naturalist 78 S.J. Wolbert, A.S. Zellner, and H.P. Whidden 2014 Vol. 21, No. 1 Figure 2. Relationship between hourly bat activity and elevation at three study areas in northeastern Pennsylvania in 2006: a) Crystal Lake, b) Hypsie Gap, and c) Hickory Run. Box = interquartile range (25%–75%), heavy band = median, whiskers = minimum and maximum values up to 1.5 times interquartile range, circles = outliers beyond 1.5 times interquartile range. Note differences in scale of y-axis among plots. Northeastern Naturalist Vol. 21, No. 1 S.J. Wolbert, A.S. Zellner, and H.P. Whidden 2014 79 was greater at the upper-elevation sampling sites (P = 0.021). Insect biomass was negatively associated with bat activity (P < 0.001). On 15 nights of sampling (14 nights at Hickory Run and 1 at Crystal Lake), there was a complete temperature inversion, i.e., temperatures increased consistently with elevation. Our regression model assessing the effect of temperature inversions included hourly data from 14 nights with inversions and 6 nights without inversions at the Hickory Run upper-elevation (579 m) site. Bat passes and insect biomass were response variables, and hour and inversion, plus the interaction between hour and inversion (hour:inversion), were predictor variables. The interaction between temperature inversion and hour was not significant for bat activity (P = 0.10) or insect biomass (P = 0.88). However, the highest levels of bat activity occurred on nights with a temperature inversion, and there was a non-significant trend towards increasing bat activity at higher elevation on nights with a temperature inversion (Fig. 4a). In contrast, insect biomass increased with hour on nights both with and without a temperature inversion (Fig. 4b). Discussion We found considerable variation in bat activity across our three study areas in northeastern Pennsylvania. The Hickory Run study area had significantly more bat activity than Crystal Lake, and Crystal Lake had significantly more than Hypsie Gap. These results are consistent with previous acoustic monitoring studies that have documented extensive spatial and temporal variation in bat activity (e.g., Table 2. Estimated regression coefficients, standard error, 95% confidence interval, z-values, and Pvalues for the negative binomial regression model predicting the number of bat passes at 3 sites in northeastern Pennsylvania. Study area, elevation, and season were categorical variables, with Crystal Lake, Lower Elevation, and Pre-volant Season serving as reference levels. The model included 238 observations (observation = complete data for 5 hours of sampling at an elevation). Predictor Estimate SE 2.5 % 97.5% z-value P-value Intercept 3.288 0.412 2.384 4.233 7.979 <0.001 Hypsie Gap study area -2.229 0.299 -2.829 -1.622 -7.462 <0.001 Hickory Run study area 2.366 0.293 1.781 2.954 8.085 <0.001 Middle elevation -0.496 0.559 -1.745 0.740 -0.887 0.375 Upper elevation -1.638 0.598 -2.930 -0.343 -2.740 0.006 Volant season -0.838 0.158 -1.183 -0.493 -5.292 <0.001 Temperature 0.175 0.023 0.120 0.231 7.670 <0.001 Insect Biomass -0.207 0.048 -0.319 -0.089 -4.299 <0.001 Hypsie Gap:Middle elevation 0.811 0.420 -0.038 1.657 1.931 0.053 Hickory Run:Middle elevation -2.356 0.410 -3.178 -1.536 -5.747 <0.001 Hypsie Gap:Upper elevation 0.997 0.423 0.172 1.821 2.355 0.019 Hickory Run:Upper elevation -2.938 0.429 -3.803 -2.065 -6.850 <0.001 Middle elevation:Temperature -0.022 0.032 -0.096 0.053 -0.682 0.495 Upper elevation:Temperature 0.078 0.034 -0.003 0.157 2.304 0.021 Northeastern Naturalist 80 S.J. Wolbert, A.S. Zellner, and H.P. Whidden 2014 Vol. 21, No. 1 Hayes 1997). We also found significantly less activity at the upper-elevation sites, and interaction among study areas and elevation, with a significant positive effect at the upper-elevation Hypsie Gap site and a highly significant negative effect for the Figure 3. Relationship between air temperature and a) bat activity and b) insect biomass at three study areas in Pennsylvania. Each circle represents an hourly data point. Northeastern Naturalist Vol. 21, No. 1 S.J. Wolbert, A.S. Zellner, and H.P. Whidden 2014 81 Figure 4. Nocturnal variation (hours after official sunset) in a) bat activity and b) insect biomass at Hickory Run Upper Elevation sampling site. Open circles represent nights with a temperature inversion (n = 14), solid triangles represent nights without a temperature inversion (n = 6). Solid line = regression line for nights with temperature inversion, dashed line = regression line for nights without inversion. Northeastern Naturalist 82 S.J. Wolbert, A.S. Zellner, and H.P. Whidden 2014 Vol. 21, No. 1 Hickory Run middle- and upper-elevation sampling sites. The exceptionally high levels of bat activity recorded at the lower-elevation Hickory Run sampling site may be due to its location adjacent to a medium-sized perennial stream (Hickory Run) where we regularly observed bats foraging on emerging insects. Bat activity showed a significant positive association with temperature, with activity beginning to increase at approximately 10 °C. Hourly samples with high levels of bat activity were recorded through approximately 18 °C, after which activity declined, although many samples at all temperatures had low levels of activity. These results are consistent with previous studies. For example, Rydell et al. (1996) recorded no observations of bats when the temperature was below 10 °C, and Reynolds (2006) found no detectable migration activity when the mean nightly temperature was below 10.5 °C. In addition, Agosta et al. (2005) noted a similar wide range of variability in bat activity and capture success above the 10 °C threshold temperature. Many state and federal agencies require that bat surveys be conducted when temperatures are above 10 °C, and our data support this cutoff as a reasonable threshold for assessments of bat activity. The interaction between elevation and temperature was significant, indicating that the effect of temperature on bat activity depended in part upon elevation, with temperature having a significantly greater effect on bat activity at upper-elevation sites. Insect biomass began to increase at approximately 8 °C and remained high in many hourly samples until the temperature reached 28 °C. This positive association of insect biomass with temperature is consistent with the general conclusions of previous research (e.g., Mellanby 1939, Raimondo et al. 2004), and Taylor (1963) similarly found very little insect activity under 10 °C. Surprisingly, bat activity had a highly significant but negative relationship with insect biomass (P = 0.021). This result contradicts Hayes (1997), who found a highly significant positive correlation between insect biomass and bat activity at two study areas (rs= 0.481, n = 70, P < 0.001, and rs= 0.388, n = 73, P < 0.001). However, Grindal and Brigham (1999) found that bat activity was not significantly correlated with insect biomass (r = -0.03, P > 0.5), and Wickramasinghe et al. (2004) found a very weak and nonsignificant relationship between bat activity and insect biomass (rs = 0.175, P = 0.235). A possible explanation for the negative relationship between bat activity and insect biomass in our model is that the highest levels of bat activity were all recorded at temperatures between 10–18 °C, whereas the highest levels of insect biomass were collected above 18 °C, and in contrast to bat activity, hourly samples with high levels of insect biomass were common at temperatures up through 28 °C. Contrary to expectations, bat activity in the volant period was significantly less than during the pre-volant period. This finding may be because the volant period included samples from as late as early November, when temperatures were lower and some summer resident bats may have entered hibernation or migrated south. We documented a series of peaks in bat activity starting in mid-August that corresponded with the beginning of fall migration activity and movements to hibernacula in mid- to late August (Merritt 1987, Whitaker and Hamilton 1998). This period of high activity also corresponds to peaks in bat mortality at wind facilities. Northeastern Naturalist Vol. 21, No. 1 S.J. Wolbert, A.S. Zellner, and H.P. Whidden 2014 83 For example, Johnson et al. (2003) found peaks of bat mortality starting in mid- to late July and continuing through mid- to late August. Hayes (1997) and Erickson and West (2002) documented the same general seasonal trends in bat activity, with increased activity in late August. Our analysis of bat-activity levels in response to temperature inversions did not support the hypothesis of increased higher-elevation bat activity on nights with temperature inversions (P = 0.1). However, due to the failure of a data logger at the upper-elevation Hickory Run sampling site from 13 September–7 November 2006, our sample sizes for this site were relatively small (only 14 nights with inversions and 6 nights without inversions). As noted above, the highest levels of bat activity all occurred on nights with a temperature inversion and there was some suggestion of a trend towards increasing bat activity at higher elevation on nights with a temperature inversion. Additional research may reveal a significant effect of temperature inversions on bat activity. However, our results suggest that any association between increased bat activity and temperature inversions is not due to increased availability of insect prey since temperature inversions did not lead to increased insect biomass at higher elevations, and bat activity had a negative association with insect biomass. Temperature, not prey availability, is more likely to be the primary factor responsible for any shifts in bat activity during such inversions. Acknowledgments We would like to acknowledge Ryan McLaughlin, Mario DaSilva, Jr., Shannon Williams, Anthony McBride, Kevin O’Driscoll, and Amie D’Angelo for help with field work, Gregory Turner, Calvin Butchkoski, and Justin Vreeland for help with study design and planning, and Eugenia Skirta for help with statistics. Tom LaDuke, Terry Master, and Matt Wallace provided helpful comments on earlier drafts of this manuscript. Peter Paton and two anonymous reviewers provided insightful and thorough reviews that greatly improved the final manuscript, and we very much appreciate their assistance. We would also like to acknowledge the Pennsylvania Game Commission, Bat Conservation International, the Pennsylvania Department of Conservation and Natural Resources, and East Stroudsburg University for funding and support. Literature Cited Agosta, S.J., D. Morton, B.D. Marsh, and K.M. Kuhn. 2005. Nightly, seasonal, and yearly patterns of bat activity at night roosts in the central Appalachians. Journal of Mammalogy 86:1210–1219. Arbuthnott, D., and R.M. Brigham. 2007. The influence of a local temperature inversion on the foraging behavior of Big Brown Bats, Eptesicus fuscus. Acta Chiropterologica 9:193–201. Arnett, E.B, W.K. Brown, W.P. Erickson, J.K. Fiedler, B.L. Hamilton, T.H. Henry, A. Jain, G.D. Johnson, J. Kerns, R.R. Koford, C.P. Nicholson, T.J. O’Connell, M.D. Piorkowski, and R.D. Tankersley, Jr. 2008. Patterns of bat fatalities at wind-energy facilities in North America. Journal of Wildlife Management 72:61–78. Barclay, R.M.R. 1991. Population structure of temperate zone insectivorous bats in relation to foraging behavior and energy demand. Journal of Animal Ecology 60:165–178. Northeastern Naturalist 84 S.J. Wolbert, A.S. Zellner, and H.P. Whidden 2014 Vol. 21, No. 1 Bates, R.L., and J.A. Jackson. 1984. Dictionary of Geological Terms, 3rd Edition. Doubleday, New York, NY. 571 pp. Carter, T.C., M.A. Menzel, B.R. Chapman, and K.V. Miller. 2004. Partitioning of food resources by syntopic Eastern Red (Lasiurus borealis), Seminole (L. seminolus), and Evening (Nycticeius humeralis) Bats. American Midland Naturalist 151:186–191. Crawley, M.J. 2013. The R Book, 2nd Edition. John Wiley and Sons, Chichester, West Sussex, UK. 1051 pp. Cryan, P.M., and R.M.R. Barclay. 2009. Causes of bat fatalities at wind turbines: Hypotheses and predictions. Journal of Mammalogy 90:1330–1340. Erickson, J.L., and S.D. West. 2002. The influence of regional climate and nightly weather conditions on activity patterns of insectivorous bats. Acta Chiropterologica 4:17–24. Field, A., J. Miles, and Z. Field. 2012. Discovering Statistics Using R. Sage Publications, Thousand Oaks, CA. 957 pp. Fox, J., and G. Monette. 1992. Generalized collinearity diagnostics. Journal of the American Statistical Association 87:178–183. Fraser, E.E., L.P. McGuire, J.L. Eger, F.J. Longstaffe, and M.B. Fenton. 2012. Evidence of latitudinal migration in Tri-colored Bats, Perimyotis subflavus. PLoS ONE 7:e31419. doi:10.1371/journal.pone.0031419. Gannon, W.L, R.E. Sherwin, and S. Haymond. 2003. On the importance of articulating assumptions when conducting acoustic studies of habitat used by bats. Wildlife Society Bulletin 31:45–61. Grindal, S.D., and R.M. Brigham. 1999. Impacts of forest harvesting on habitat use by foraging insectivorous bats at different scales. Ecoscience 6:25–34. Hayes, J.P. 1997. Temporal variation in activity of bats and the design of echolocationmonitoring studies. Journal of Mammalogy 78:514–524. Johnson, G.D., W.P. Erickson, M.D. Strickland, M.F. Shepard, and D.A. Shepard. 2003. Mortality of bats at a large-scale wind-power development at Buffalo Ridge, Minnesota. American Midland Naturalist 150:332–342. Johnson, G.D., M.K. Perlik, W.P. Erickson, and M.D. Strickland. 2004. Bat activity, composition, and collision mortality at a large wind plant in Minnesota. Wildlife Society Bulletin 32:1278–1288. Kerns, J., and P. Kerlinger. 2004. A study of bird and bat collision fatalities at the Mountaineer Wind Energy Center, Tucker County, West Virginia: Annual report for 2003. Prepared for FPL Energy and Mountaineer Wind Energy Center Technical Review Committee by Curry and Kerlinger, LLC, McLean, VA. Kunz, T.H., E.B. Arnett, B.M. Cooper, W.P. Erickson, R.P. Larkin, T. Mabee, M.L. Morrison, M.D. Strickland, and J.M. Szewczak. 2007a. Assessing impacts of wind-energy development on nocturnally active birds and bats: A guidance document. Journal of Wildlife Management 71:2449–2486. Kunz, T.H., E.B. Arnett, W.P. Erickson, A.R. Hoar, G.D. Johnson, R.P. Larkin, M.D. Strickland, R.W. Thresher, and M.P. Tuttle. 2007b. Ecological impacts of wind-energy development on bats: Questions, research needs, and hypotheses. Frontiers in Ecology and the Environment 5:315–324. Kurta, A., and J.O. Whitaker, Jr. 1998. Diet of the endangered Indiana Bat (Myotis sodalis) on the northern edge of its range. American Midland Naturalist 140:280–286. Lacki, M.J., L.S. Burford, and J.O. Whitaker, Jr. 1995. Food habits of Gray Bats in Kentucky. Journal of Mammalogy 76:1256–1259. Mellanby, K. 1939. Low temperature and insect activity. Proceedings of the Royal Society of London, Series B, Biological Sciences 127:473–487. Northeastern Naturalist Vol. 21, No. 1 S.J. Wolbert, A.S. Zellner, and H.P. Whidden 2014 85 Merritt, J.F. 1987. Guide to the Mammals of Pennsylvania. University of Pittsburgh Press, Pittsburgh, PA. 408 pp. Moosman, P.R., Jr., H.H. Thomas, and J.P. Veilleux. 2012. Diet of the widespread insectivorous bats Eptesicus fuscus and Myotis lucifugus relative to climate and richness of bat communities. Journal of Mammalogy 93:491–496. R Core Team. 2013. R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. Available online at http://www.R-project. org/. Accessed 16 May 2013. Raimondo, S., J.S. Strazanac, and L. Butler. 2004. Comparison of sampling techniques used in studying Lepidoptera population dynamics. Environmental Entomology 33:418–425. Reynolds, D.S. 2006. Monitoring the potential impact of a wind-development site on bats in the Northeast. Journal of Wildlife Management 70:1219–1227. Rydell, J., A. Entwistle, and P.A. Racey. 1996. Timing of foraging flights of three species of bats in relation to insect activity and predation risk. Oikos 76:243–252. Sherwin, R.E., W.L. Gannon, and S. Haymond. 2000. The efficacy of acoustic techniques to infer differential use of habitat by bats. Acta Chiropterologica 2:145–153. Taylor, L.R. 1963. Analysis of the effect of temperature on insects in flight. Journal of Animal Ecology 32:99–117. Triplehorn, C.A., and N.F. Johnson. 2005. Borror and Delong’s Introduction to the Study of Insects. Seventh Edition. Thomson Brooks/Cole, Belmont, CA. 864 pp. Weller, T.J., and C.J. Zabel. 2002. Variation in bat detectors due to detector orientation in a forest. Wildlife Society Bulletin 30:922–930. Whitaker, J.O., Jr. 2004. Prey selection in a temperate zone insectivorous bat community. Journal of Mammalogy 85:460–469. Whitaker, J.O., Jr., and W.J. Hamilton, Jr. 1998. Mammals of the Eastern United States, 3rd Edition. Cornell University Press, Ithaca, NY. 583 pp. Wickramasinghe, L.P., S. Harris, G. Jones, and N.V. Jennings. 2004. Abundance and species richness of nocturnal insects on organic and conventional farms: Effects of agriculture intensification on bat foraging. Conservation Biology 18:12 83–1292.