Observed Resiliency of Little Brown Myotis to Long-Term White-Nose Syndrome Exposure

White-nose syndrome (WNS) is a disease that has decimated bats in eastern North America and has steadily been spreading into susceptible bat populations across the continent. The WNS epizootic is thought to have originated in Europe (Leopardi et al. 2015; Campana et al. 2017; Drees et al. 2017) and was first detected in the United States in 2006 at Howe's Cave in Schoharie County, New York (Blehert et al. 2009). White-nose syndrome has since spread to 32 states and five Canadian provinces (U.S. Fish and Wildlife Service [USFWS] 2017). In addition, evidence of the causative agent of the disease, the novel psychrophilic fungus Pseudogymnoascus destructans (Blehert et al. 2009; Gargas et al. 2009; Minnis and Lindner 2013), has been detected on bats as far south and west as Mississippi and Washington, respectively (USFWS 2017). To date, WNS has severely impacted some of the most common temperate North American bat species, including little brown myotis Myotis lucifugus. This species has experienced an approximate 90% decline at hibernation sites in the Northeast (Frick et al. 2010; Turner et al. 2011; Langwig et al. 2017). Similar declines in little brown myotis summer colonies and activity on the landscape have also been noted throughout the Northeast (Ford et al. 2011; C. Herzog, A. Bennett, and S. von Oettingen, personal communication), Mid-Atlantic, and Midwest regions (Francl et al. 2012; Pettit and O'Keefe 2017), with many colonies and populations now extirpated. However, despite extensive declines, some localized populations of little brown myotis are persisting (Dobony et al. 2011; Reichard et al. 2014).

Examination of the mechanisms that may be leading to long-term survival and persistence of WNS-impacted populations is important for understanding the overall dynamics of both bat and disease management. The continued persistence and long-term survival of WNS-impacted individuals will likely be key components in overall species recovery, especially if there is any kind of heritable resiliency or resistance to the fungus or the disease as a whole. Even if no genetic component is passed on to offspring, surviving individuals may still be important in population recovery through continued contributions to reproductive output at maternity colonies, passing on beneficial learned behaviors to cope with P. destructans or WNS, or in other advantageous ways presently unknown. Currently, it is unclear what mechanisms are contributing to the continued persistence of little brown myotis in some areas. Leading hypotheses have focused on whether there is an immune-mediated response, whether some bats are developing a resistance or tolerance to the disease and fungus, or whether both responses may be occurring. Currently, the immune response does not seem to be contributing to this persistence (Johnson et al. 2015; Lilley et al. 2017). Although there is evidence that some bats could either possess or have developed some kind of resistance or tolerance (Frick et al. 2017; Langwig et al. 2017), comparative genomics of the genus Pseudomgymnoascus suggests P. destructans has actually evolved to live on bats (Palmer et al. 2018). If this is the case, it is unlikely that susceptible species in suitable habitats will be able to avoid fungal colonization and WNS infection. Perhaps the surviving populations indicate there is a similar equilibrium now being established in North America as Zukal et al. (2016) suggested may exist for bats and P. destructans in Europe and Palearctic Asia, where there is no contemporary evidence of mass mortality comparable to what is observed in North America (Martinkova et al. 2010; Zukal et al. 2016). Still, other researchers have suggested that density-dependent disease responses or different behaviors in different geographic ranges may allow some bats to remain (Langwig et al. 2012; Bernard and McCracken 2017).

Persistence of species heavily impacted by disease may not be due to a single, straightforward cause; rather, it could be a result of an interaction of multiple biologic, social, and environmental factors (Stephen 2014). Consequently, it can be difficult to directly measure or deduce exact mechanisms underlying persistence of a species. Furthermore, factors contributing to the resilience of a species, or the ability to maintain population integrity despite disturbance (Holling 1973; Redford et al. 2011) such as disease, can be manifold. It is likely that these interactions take place over the lifetime of a bat, and short-term studies may not fully document accurate responses to disease. Therefore, we monitored a little brown myotis summer maternity colony in New York that has been impacted by WNS for multiple years subsequent to initial WNS exposure. To contribute to the accumulation of WNS-impacted bat population knowledge in the Northeast and attempt to identify mechanisms that may be responsible for this colony's continued existence post-WNS exposure, our objectives were to conduct long-term monitoring and annual census of the colony, obtain demographic data, monitor bats for evidence of continued P. destructans and WNS exposure, and develop survival and recapture models over 10 y post-WNS exposure. We hypothesized that the bat population at this colony would either continue to decline until extirpation or a segment of the colony would persist and gradually increase over time.

We conducted our study at Fort Drum, an >43,000-ha U.S. Army Installation in Jefferson and Lewis counties in northern New York (44°00′N, 75°49′W). Fort Drum included a cantonment area, main impact area, airfield, and 18 training areas. Approximately 57% of the installation (∼25,000 ha) was forested with mature northern hardwood types of sugar maple Acer saccharum, American beech Fagus grandifolia, and white ash Fraxinus americana, with softwood associates of white pine Pinus strobus and Eastern hemlock Tsuga canadensis found throughout. Early successional habitat was dominated by red maple Acer rubrum, gray birch Betula populifolia, and quaking aspen Populus tremuloides. Beaver ponds, small lakes, wet meadows, and other wetland systems made up approximately 20% of the land cover (∼8,000 ha). Extensive development was concentrated within the cantonment area and certain firing ranges encircling the main impact area; only scattered training facilities and other structures were found throughout the relatively undeveloped maneuver areas in the remainder of the installation. Elevation ranged from 125 to 278 m.

In 2003, monitoring of bat community demography and spatial and temporal distribution was initiated. During that time, a relatively large (>1,000 individuals) summer maternity colony of little brown myotis was discovered in the historic 19th century LeRay Mansion. Evidence of bat activity at the mansion suggested the colony had roosted in the structure for >20 y. The long-term occupancy of the mansion led to guano accumulation in wall cavities and inaccessible attic space and raised concerns for human health and structural integrity. In response, a bat house was installed approximately 200 m from the mansion in 2004, and bat exclusion and relocation efforts were initiated. Individual bats were documented using both roosts interchangeably throughout the exclusion period; however, by summer 2008 approximately 70% of the bats had moved into the bat house. White-nose syndrome was first suspected to be affecting bats in the colony in summer 2008, and an intensive banding program was initiated in summer 2009 to establish an effective way to monitor the potential impacts from WNS on survival, reproduction, and recruitment (Dobony et al. 2011). Only a small number of bats (<20) were known to occasionally use the mansion post-WNS; therefore, exclusion was considered complete by summer 2009 and we focused monitoring and sampling efforts primarily on the bat house. Throughout the study, the little brown myotis colony at Fort Drum was one of less than a dozen monitored summer colonies of any numerical significance (>100 individuals) remaining in New York and New England (C. Herzog and A. Bennett, personal communication).

We conducted postvolancy emergence counts of little brown myotis at both the mansion and nearby bat house in Fort Drum's cantonment area in 2006 and 2008 and pre- and postvolancy counts at the bat house from 2009 to 2017. We considered counts conducted in May or early to mid-June as representing adults because juvenile bats were not typically volant at this time (C. Dobony, unpublished data). We assumed that counts after the first week of July represented both adults and postvolancy juveniles. We recorded the timing and number of bats exiting the bat house. We began counts approximately 30 min before sunset and continued until either bats finished exiting or darkness precluded accurate counting. The bottom of the bat house was approximately 4 m off the ground and the roost was open underneath. This allowed for easy illumination and observation to determine whether any individuals remained in the bat house after counting ended. It would have been ideal to combine counts of any remaining bats within the house with counts of all bats observed exiting for a more accurate representation of the total size of the colony; however, sometimes this was not possible. In some cases, too many bats were clustered into a baffle within the house to distinguish individuals, or bats were moving back and forth within the house (as was typically the case with young of the year bats). Therefore, we only included bats observed exiting the house to estimate numbers of adults and juveniles and total colony size. This provided a less accurate, but more repeatable, consistent, and precise estimate.

We captured bats from the bat house during 22 capture events from May to August 2009–2017 by using three methods: 1) we funneled bats emerging from the roost into a double-frame harp trap (183 × 183 cm; Bat Conservation and Management, Carlisle, PA) or low-visibility mist nets (Avinet Inc., Dryden, NY) by using coarse nylon bird netting (E. I. du Pont de Nemours and Company, Wilmington, DE); 2) we attached 1.5-mil plastic sheeting to the underside of the roost and used it to funnel bats directly into holding cages or to personnel; and 3) we placed mist nets close (5–75 m) to the bat house along known exit corridors to woods and water. We determined species, sex, age (Brunet-Rossinni and Wilkinson 2009), and reproductive condition (i.e., pregnant, lactating, or postlactating; Racey 2009). We marked each bat with a split-ring, lipped, alloy band with a unique identification number (Porzana Ltd., Icklesham, UK). We then assessed the bats for evidence of recent past infection from WNS by visually inspecting wing membranes for damage. We ranked the condition numerically from 0 to 3, where 0 indicated no evidence of damage and 3 indicated the highest level of damage (Reichard and Kunz 2009). Currently, there are no other minimally invasive, nonlethal methods suitable for use for determining recent past WNS infection when sampling bats outside of hibernation season in early-to-late summer. We used wing damage index (WDI) as a proxy indicator of recent past infection, not a diagnostic tool. Therefore, a WDI of 0 does not indicate that the bat had no recent infection; rather, it was undetermined based on the condition of the wings. Conversely, classic damage from WNS leading to a WDI of 1, 2, or 3 did indicate recent past infection. Although bat wing membranes can be damaged in many ways, we had sufficient experience assessing wing conditions to distinguish between recent past WNS infection (i.e., lesions, splotching, scarring, or necrotic tissue) and other physical damage. In northern New York, wing damage associated with previous WNS infections can be consistently documented on bats at summer roosts from late April through mid-July (Dobony et al. 2011).

We sampled a subset of bats in May 2010 (n = 15), June 2010 (n = 18), July 2010 (n = 24), and July 2014 (n = 12) to determine presence of P. destructans. We collected 3-mm sterile biopsy punches (Miltex, Inc., York, PA) from bats during each sampling event in 2010. If we did not immediately freeze the punches, we placed them in 95% ethanol or in a lysis buffer and then froze them for later analysis. We also collected swab samples from bats in 2014. We used single-use sterile swabs (Fisherbrand polyester-tipped applicators; Fisher Scientific Company LLC, Pittsburgh, PA) moistened in sterile water and then rolled them across the ventral and dorsal surfaces of each wing 10 times. We placed the swabs in 1.8-mL storage vials filled with RNAlater, transported them in a dry shipper, and stored them at −80°C for later analysis.

We also sampled for the presence of the fungus in the structure of the bat house in March 2011 (n = 36), September 2011 (n = 20), and November 2012 (n = 20) and for both the presence (n = 65) and viability (n = 60) of the fungus from March to November 2013. We collected samples only when bats were not physically present within the structure (i.e., before their arrival in spring, after their departure in fall to hibernation sites, or after they had left for the evening while they were using the bat house). During each sampling event, we moistened a Fisherbrand single-use sterile swab in distilled water and then rubbed back and forth six times across the width of each roosting baffle at the lower edge (i.e., where bats first landed to climb into the structure) and at the middle and upper edges (i.e., where bats mainly roosted). We then combined all swabs collected in each section (i.e., lower, middle, upper) within one sample. We preserved the swab samples in distilled water and froze them until later extraction or culture. We extracted genomic DNA with the DNeasy Blood and Tissue kit (Qiagen, Hilden, Germany) by using the manufacturer's supplementary protocol (i.e., DY13-Purification of total DNA from yeast using the DNeasy Blood and Tissue kit), as described in Shuey et al. (2014). We analyzed all bat swab, biopsy, and structure swab samples for the presence of P. destructans DNA through real-time PCR (Muller et al. 2013), and used a cycle threshold cut-off value of 40 cycles to confirm the presence of the fungus (Janicki et al. 2015; Langwig et al. 2015).

We cultured swabs collected in 2013 in Sabouraud dextrose broth and incubated them at 10°C (n = 60) for 3 wk. We then removed 200 μL of broth, plated it on Sabouraud dextrose agar, and distributed it by using a sterile cell spreader. After 3 wk, we isolated potential P. destructans colonies and replated them on Sabouraud dextrose agar. We incubated these plates at 10°C for 6 wk to allow for ample growth. We then extracted a 1-μL loop of culture and tested for the fungus. We handled each bat, swab, and culture sample with new disposable latex or nitrile gloves, and decontaminated or sterilized all equipment between sampling to avoid cross-contamination.

We calculated reproductive rates by determining the percentage of captured adult females showing signs of reproduction (i.e., pregnant, lactating, or postlactating). For individuals that we captured multiple times within a given year, we counted their reproductive status only once as part of the analysis. We could not always reliably determine reproductive status for some females; for example, for spring and early summer captures, it was sometimes difficult to tell whether a female was pregnant through palpation. Also, some females escaped before we could determine reproductive status. In either case, we eliminated these females from the analysis.

In program MARK, we used Cormack–Jolly–Seber models to estimate annual survival rates of little brown myotis. Although the assumption of a closed population often is violated by bats, colonies in permanent structures show stronger fidelity than those in more ephemeral structures (Lewis 1995; Norquay et al. 2013), thus minimizing and stabilizing biases. Sample sizes of male little brown myotis were too small for reliable survival rate estimations; typically, only young of the year males are found in little brown myotis maternity colonies. Consequently, we only included females in our models. Because the number of surveys that took place each year varied (n = 1–8), we specified time intervals in MARK and estimated survival on a weekly basis. Within a parameter index matrix in MARK, we condensed weekly parameter estimates (Ŝ_w) to a single parameter within sample years (2009–2016). We used the formula Ŝ_w⁵² = Ŝ_a to obtain annual survival estimates (Ŝ_a). We adjusted weekly survival rate variances to annual survival rate variances following Powell (2007). We could not estimate survival rate for 2016 because it was the last survey year (i.e., survival and recapture rates were confounded).

To determine whether survival rates were similar between juvenile and adult little brown myotis, we used quasi-Akaike's Information Criterion (QAIC_c) for small sample sizes to assess support between a model incorporating age structure (i.e., juveniles had different survival rate parameters their first year, but they had the same survival rate parameters as adults their second year) and a model without age structure (i.e., juveniles and adults had the same survival rate parameters within a given year). In a subsequent analysis, we used Cormack–Jolly–Seber models and QAIC_c to examine differences among adult little brown myotis that either exhibited or did not exhibit signs of WNS exposure based on wing damage score (i.e., whether they had a wing damage index of 0 or ≥1) and that either showed signs of reproductive activity or not (signs of pregnancy or lactation, or not). We could not include juveniles in the latter model because those that were not recaptured as adults could not unequivocally be placed in a reproductive condition category. We categorized and treated recapture rates the same as survival rates, except recapture rates had their own parameter estimates. We considered candidate models within 2 QAIC_c of the best candidate model as competing models. We measured goodness of fit of the best model using a median ĉ test in MARK. A median ĉ value <3.0 indicates data are not overdispersed and are a reasonable fit to the model. We used the ĉ value to adjust survival rate variances.

The highest colony count during our study was 1,200; this count included both adults and postvolancy juveniles in 2008 (Table 1). Exit counts were lowest in 2010, when 145 individuals (adults and juveniles) were counted postvolancy. This represented a total decline in colony size of approximately 88% from 2008 levels (Table 1). Since 2010, exit counts of all bats at the colony have increased annually to 256 individuals in 2017, or about an average of 14 bats per year (range = 3–48). From 2010 to 2014, the numbers of adult females returning to the colony stabilized (mean = 94, range = 84–101) and then increased after 2014 (mean = 132, range = 108–166; Table 1).

Between 2009 and 2017, we captured 727 bats, including recaptures (401 adult females, 2 adult males, 147 juvenile females, 150 juvenile males, 27 unknown-age females; Table 2). This represented 575 total female encounters and 152 total male encounters; we defined an encounter as any time we captured an individual bat during monitoring. We encountered 539 individual bats across all years. We initially captured and banded 534 bats at the bat house colony (219 adult females, 1 adult male, 143 juvenile females, 144 juvenile males, 27 unknown-age females; Table S1, Supplemental Material). There were also four bats (two adult females, one juvenile female, one juvenile male) initially captured and banded at other locations on Fort Drum that we later encountered at the bat house. One additional adult female escaped before we could determine where she originated. The juvenile female first encountered away from the bat house was originally banded in 2010 at a suspected maternity site 30.2 km from the bat house and was subsequently captured at the bat house in 2012, 2013, and 2015.

We recaptured 192 bats across all years, and one bat escaped before recapture status could be determined. Of the 192 recaptures, we recaptured 38 only within the same sampling year, 150 were from 98 female bats we captured more than once across multiple years, and the remaining recaptures were the 4 bats we captured initially during mist netting on other areas of Fort Drum during other research projects. Of these 98 recaptured bats, 82 survived at least 2 y postcapture; 53 bats survived at least 3 y; 31 survived at least 4 y; 17 survived at least 5 y; and 3 survived at least 6 y. We captured 39 of the 98 recaptured bats as juveniles. It should be noted that recapture information (i.e., rates and length of survival and persistence) reported are a minimum number. We observed some banded bats remaining in the bat house and not emerging during the sampling event. This resulted in a varying percentage on any given year that remained unavailable for recapture determination during our study.

The majority (91%, 297/328; Table 2) of adult female bats across years showed evidence of reproduction. Ninety-three percent (173/186; Table S2, Supplemental Material) of encounters with the 98 individual recaptured female bats showed evidence of reproduction. We estimated an annual average reproductive rate of 0.92 (range = 0.77–1.00) for all adult females, and recaptured adult females had a similar average annual rate of 0.95 (range = 0.85–1.00; Table 2). Almost all (97%, 38/39) female bats first encountered as juveniles had evidence of reproduction at least once during the study, and 90% (27/30; Table S2, Supplemental Material) of juvenile females recaptured at 1 y old were reproductive.

Wing damage from WNS was variable across years and was dependent on the timing of the sampling event (Table S3, Supplemental Material); higher average wing scores were measured earlier within a sampling season. We noted visible damage for the majority of encounters with adult female bats throughout the entire sampling season, and overall 61% (244/397) of WDI scores were either 1 (n = 166), 2 (n = 66), or 3 (n = 12). Most recaptured adult bats also continued to show evidence of infection, with encounters in May and June having the most visible damage. Over the course of our study, 83% (77/93; Table S2, Supplemental Material) of bats recaptured across multiple years exhibited a WDI ≥ 1 at least once. Only 6% (18/297; Table S3, Supplemental Material) of juveniles exhibited wing damage (all with a WDI of 1), and in all cases we attributed the damage to juveniles to physical injuries (e.g., puncture or tears with no necrotic tissue) and not damage from WNS. We detected P. destructans by PCR on the majority of bat biopsy (32/57) and bat swab samples (7/12) collected in 2010 and 2014, respectively (Table 3). We also detected the fungus on 67% (94/141; Table 3) of all samples collected from the interior of the bat house in 2011, 2012, and 2013 and documented it in the structure during the time of year when bats were present or not. We detected viable P. destructans in only 2 of 60 samples (Table 3); however, this is the first time that we are aware of that viable cultures have been grown from the wooden interior of a bat house during the maternity season.

Best candidate models indicated that annual survival rates were similar between juvenile and adult female little brown myotis (Table 4). Survival rates of adult female little brown myotis were not related to evidence of recent WNS exposure (i.e., wing score index values) rates. Reproductive status also was not related to survival rates (Table 5). The model without age structure had a ĉ value of 1.158 (SE = 0.028), indicating data were not overdispersed and were a good fit to our data. Similarly, the model indicating WNS and reproductive status was not important to survival estimates had a ĉ value of 1.047 (SE = 0.000), indicating data were not overdispersed and were a good fit to our data. However, survival rate estimates were unreliable for 2011 to 2012 and 2015 to 2016. Annual survival rates ranged from 41.0% from 2009 to 2010 to 86.5% from 2012 to 2013. Weekly recapture rates were not reliable for 2016; rates ranged from 3.2% in 2012 to 40.8% in 2014 (Table 6).

Although WNS caused a substantial reduction in little brown myotis numbers on Fort Drum, it did not cause a local extirpation of the bat house maternity colony as has occurred in many areas across the Northeast (C. Herzog, A. Bennett, and S. von Oettingen, personal communication) and as Dobony et al. (2011) predicted based on initial evidence collected in 2008–2010. Indeed, due to the extensive declines in the Fort Drum colony in years immediately post-WNS, and evidence from other ongoing studies on Fort Drum (Ford et al. 2011; Winhold et al. 2011), Dobony et al. (2011) suggested that a localized extirpation of myotine bats from the entire Fort Drum landscape was probable. In addition, Frick et al. (2010) predicted regional extinction of little brown myotis was likely by 2026. It now seems that these conclusions may have been premature for little brown myotis in the Northeast. Pettit and O'Keefe (2017) documented similar declines of myotine bats based on summer surveys and suggested that WNS may have caused the local extinction of little brown myotis in Indiana. Although Pettit and O'Keefe (2017) monitored bats for 13 y (2002–2014), only 3 y occurred after suspected WNS infection. As illustrated in Dobony et al. (2011), even after 3 y of monitoring post-WNS at Fort Drum, the picture was not clear, nor complete. It has taken almost 10 y post-WNS monitoring to see additional changes within the Fort Drum colony. Not only has the maternity colony at Fort Drum stabilized and increased but also other surveys on Fort Drum have documented increases of little brown myotis numbers on the summer landscape after multiple years of low capture rates (Baer et al. 2016).

Unfortunately, the mechanisms leading to this persistence remain unclear. It is possible given existing evidence that surviving bats may be developing a type of tolerance or resistance to P. destructans and the effects of WNS (Zukal et al. 2016; Frick et al. 2017), although it still seems more complex than that. We suggest it is perhaps more appropriate at this time to refer to what is being observed in these remnant populations as a type of broader resiliency that could encompass a broad spectrum of mechanisms allowing some bats to continue to heal, reproduce, recover, adapt, and ultimately persist despite WNS. Although no simple “smoking gun” emerged in the parameters we examined as any likely factor leading to resiliency and persistence in this population, we suggest that perhaps any of the factors examined could play some role. We were also able to rule out what is seemingly not important when looked at individually and provide important long-term observations within a resilient WNS-impacted colony.

Little brown myotis at Fort Drum are not persisting by avoiding P. destructans and subsequent WNS infection. Initial evidence indicated that bats were likely exposed to P. destructans during fall swarming, winter hibernation season, or both, and developed at least some level of WNS infection annually (Dobony et al. 2011). Frick et al. (2017) suggested that few bats escape exposure and by early winter most of the population at P. destructans–positive hibernation sites are infected. They suggested that early winter infections could likely be explained by P. destructans persisting within the hibernation sites and immediately infecting bats as they arrived. Lorch et al. (2013), Reynolds and Barton (2014), and Hoyt et al. (2015b) determined that P. destructans could persist in the environment, even in the absence of bats. Dobony et al. (2011) initially isolated viable P. destructans on two juvenile female bats at the Fort Drum research colony in August 2009, and viable fungus was subsequently found in the structure during our sampling in May 2013. The presence of the fungus was also documented within the interior structure of the bat house throughout the entire year, both when bats were present and absent, making this the first time the fungus has been confirmed both present and viable within an artificial summer maternity roosting structure. In addition, other unpublished research at Fort Drum has documented the presence of P. destructans within the guano of the roosting bats during spring and early summer.

All of this evidence indicates that even environmental reservoirs such as summer maternity sites harbor the fungus for indeterminate lengths of time, and bats exposed to these reservoirs of P. destructans in late summer may play a larger role in transporting the fungus to uncontaminated sites and roost mates than previously thought. These summer sites could contribute to early winter infection if some bats are continually picking up and transporting viable P. destructans spores into hibernation sites or if bats are already colonized with the fungus before entering hibernation. Although testing to date has only shown the fungus to be viable in the bat house or on bats in very few samples after early summer and before hibernation, this is still concerning. Ballmann et al. (2017) reported that bats occupying hibernacula in the summer are at risk of exposure to P. destructans and may carry and disperse the fungus during the nonhibernation season, and we have now documented that this risk exists for bats at summer maternity sites as well.

Langwig et al. (2017) examined little brown myotis populations persisting at hibernation sites in eastern New York and found that although there were much lower fungal loads on average in these hibernacula vs. those where the fungus had recently invaded, the bats were not able to avoid P. destructans exposure and subsequent infection. Further corroborating these findings, Johnson et al. (2015) and Lilley et al. (2017) determined that an immune-mediated response against P. destructans and WNS is unlikely in little brown myotis. We confirmed that annual exposure and infection is still occurring, as most of the surviving (recaptured) bats in our study continued to show evidence of WNS infection and wing damage in early May when they had returned from hibernation sites. Even given this repeated exposure and infection cycle, bats began to heal in the late spring and early summer, and in many cases by the end of summer WNS damage was nearly undetectable. This has been documented on bats that have been recaptured multiple times within a single season at Fort Drum and other areas in New England (Dobony et al. 2011; Fuller et al. 2011). In the context of this population persisting, continued exposure to the fungus and evidence of recent past infection seemingly play no role in influencing annual survival rates of returning bats. It is possible that no differences were found simply because all adult females returning to the colony have experienced some level of fungus exposure or recent WNS infection regardless of the WDI we assigned. As indicated previously, we used the WDI as a proxy, not an absolute diagnostic tool. Therefore, bats with a WDI of 0 could have conceivably experienced just as much recent infection as those with a WDI of 1, 2, or 3. The bats with a WDI of 0 perhaps left hibernation sooner and were subsequently able to clear infection and heal before our sampling, or the damage they sustained was less severe, leading to quicker healing rates. Outside of wing damage too extensive for bats to fly, leave the hibernacula, and subsequently forage, it may not matter how much other lesser damage is sustained by some of these bats. The survivors may just be the survivors regardless, as we have documented recaptured bats returning with all manner of wing damage over multiple years.

It has been suggested that energetic costs of disease, healing, and compromised flight may affect reproductive rates in WNS-affected populations (Fuller et al. 2011). We found no evidence of this. Successful reproduction has been documented in bats at Fort Drum since evidence of WNS was initially observed, and during our study, reproductive rates were generally better than those Dobony et al. (2011) observed. Although reproductive rates were variable, we noted that they have rebounded to relatively stable and high values, similar to those observed at another maternity site that was extensively monitored in New Hampshire for 16 y before the WNS outbreak (Frick et al. 2009). In contrast to that study, a 1-y-old female reproductive rate of 90% in our study was much higher than that predicted by models presented by Frick et al. (2009; 23–53%). Therefore, 1-y-old females are potentially contributing more significantly to reproductive effort at the Fort Drum colony than expected. Although relatively limited recaptures of 1-y-old females temper our ability to draw strong conclusions, it is possible that this could be an evolutionary mechanism or a density-dependent response to the high mortality rates associated with WNS. If 1-y-old female bats are successful in rearing pups to volancy and recruiting them into the maternity colony, having those individuals contributing to reproductive output could lead to an increased recovery rate at the colony. Our observed range of reproductive rates from other captured bats (excluding 1 y olds) was similar to ranges observed pre-WNS (Frick et al. 2009). Although this repeated infection cycle may ultimately interfere with the resilience of bats, currently most of these females are still able to cope with this infection while still partitioning energy into healing and a normal reproductive cycle (Dobony et al. 2011). Therefore, it does not seem that WNS has caused any changes in fecundity in resilient populations such as the Fort Drum colony, and it seems unlikely that depressed reproduction is leading to a slow recovery. Rather, slow population growth is more likely a function of inherent low reproductive potential and continued variable mortality from WNS and other causes leading ultimately to lower recruitment.

Adult and juvenile survival may be a more important consideration than fecundity in little brown myotis recovery (Maslo et al. 2015). We observed no differences between adult and juvenile survival, and annual survival rates ranged between 41.0 and 86.5%, although we caution that survival rate estimates were unstable for some years, hampered by low sample sizes. Survival rates in our study were lower than what Frick et al. (2009) reported pre-WNS but similar to reported survival rates (range = 65–70%) post-WNS (Maslo et al. 2015). However, in contrast to both Frick et al. (2009) and Maslo et al. (2015), we observed more variability in the survival rate estimates. Given the lower average survival rates that we observed, rapid recovery is not likely, especially in years where survival is closer to 50.0%. Although the average survival rate (mean = 59%) seems to be high enough to result in small population increases over time, recovery will take multiple generations, even with high reproductive rates. Because little brown myotis typically only produce one pup per year, any small depression in either the reproductive rate or survival rate could result in a slowing of population growth.

Maslo et al. (2015) suggested that improved survival, and not immigration from other areas, was likely leading to stabilization at remnant hibernating colonies in WNS-affected areas. Limited sampling in our study did not allow for robust design modeling that would estimate immigration and emigration from the colony (Kendall and Nichols 1995); however, we did document one example of immigration from another maternity colony. A juvenile female that was originally banded in July 2010 as part of a separate mist-netting study, was encountered at the bat house in 2012, 2013, and 2015 approximately 30.2 km from where she was initially banded. We also captured 25 other little brown myotis, including adult females and juvenile males and females, at the 2010 mist-net site, indicating a maternity colony was located nearby. We netted the same site in 2011, and captured no bats, indicating the colony may have moved elsewhere. In 2015, we netted the site again and captured 42 little brown myotis, including adult females and juvenile males and females. Although the colony seemed to be reestablished at this location, we still found the original juvenile female at our focal study roost in 2015 and not at the site where she was initially banded.

Given this evidence, roost-switching cannot be ruled out as a possible explanation for some changes in our colony size throughout the monitoring period. There may be some interchange between the bat house colony with other unknown little brown myotis colonies on other areas of Fort Drum. However, although Norquay et al. (2013) observed this behavior over a 20-y period, a high number (∼88%) of little brown myotis still returned to their maternity sites in subsequent years. We had only one documented occurrence during this study and one other example of this behavior during an unrelated study on Fort Drum (Baer et al. 2016), either before or after WNS, over a >10-y period. So, although there is likely some interchange occurring and estimating immigration and emigration rates at the colony would permit more accurate population growth estimates resulting from recruitment and survival, it does not seem that this behavior is playing any large role in the steady increases we have been observing in recent years. Rather, given our observations of reestablishment and increases at previously suspected extirpated maternity colonies, as well as general increases across the summer landscape at Fort Drum (Baer et al. 2016), it seems that increases at our focal colony roost is part of an overall localized small-scale recovery.

Although it is encouraging to observe the continued persistence of the Fort Drum colony, it should be noted that resulting increases in local and regional hibernation counts have not been realized to the same degree (C. Herzog, personal communication). The total number and locations of hibernacula in which bats from the Fort Drum colony overwinter are unclear; however, we know significant declines within, or complete losses of, hibernating populations of little brown myotis have been documented due to WNS in New York (Turner et al. 2011). Some of these sites may have been where bats from the Fort Drum colony hibernated, not only resulting in loss of individuals from overall and local populations but also loss of genetic material. Fortunately, little brown myotis can migrate long distances and hibernate and have summer roosts in various locations, resulting in a virtual network of such locations across the landscape (Norquay et al. 2013). This not only results in high gene flow (Burns et al. 2014) but also increases the potential for reestablishing or enhancing populations that were either extirpated or significantly diminished by WNS. Conserving resilient colonies such as the one at Fort Drum will serve to maintain a small, yet potentially significant portion of the overall network of little brown myotis in the Northeast and perhaps beyond. However, additional analysis of total geographic representation of the species, resiliency, and redundancy will be important in the context of overall species recovery and returning to and maintaining a self-sustaining population (Shaffer and Stein 2000; Wolf et al. 2015).

The persistence of the colony at Fort Drum remains tenuous, and surviving bats continue to be exposed to P. destructans and infected with WNS. At some point, continued losses from WNS, other causes of mortality (including natural deaths) of the surviving bats, or a stochastic event could result in renewed decreases or subsequent extirpation of the colony. Given these concerns and the continued persistence and spread of WNS, there has been considerable attention to development of medical, fungicidal, antimicrobial, or other types of treatments specifically to stop or slow the spread of the fungus and WNS infection (Cornelison et al. 2014; Hoyt et al. 2015a; Raudabaugh and Miller 2015). Although it is essential to continue to explore practical management solutions for this devastating disease, caution must be exercised when discussing mechanisms to assist these populations, as our work has demonstrated that WNS-infected little brown myotis are capable of persisting without human intervention. If this persistence and resiliency have resulted from complex factors and behaviors developed over time (or will more fully develop given more time), human involvement could have unknown consequences and even exacerbate population declines. It is possible that refraining from intervention currently may be the best option, allowing time and surviving bats to grow the remaining population.

Please note: The Journal of Fish and Wildlife Management is not responsible for the content or functionality of any supplemental material. Queries should be directed to the corresponding author for the article.

Table S1. Capture information for little brown myotis Myotis lucifugus at a maternity colony at Fort Drum Military Installation, New York, 2009–2017.

Found at DOI: http://dx.doi.org/10.3996/102017-JFWM-080.S1 (44 KB XLSX).

Table S2. Recapture summary with wing condition scores of individual female little brown myotis Myotis lucifugus monitored at a maternity colony at Fort Drum Military Installation, New York, 2009–2017.

Found at DOI: http://dx.doi.org/10.3996/102017-JFWM-080.S2 (18 KB XLSX).

Table S3. Number of little brown myotis Myotis lucifugus captured by sampling period, age class, and wing condition score, from a maternity colony at Fort Drum Military Installation, New York, 2009–2017.

Found at DOI: http://dx.doi.org/10.3996/102017-JFWM-080.S3 (15 KB XLSX).

Reference S1. Baer Z, Hawkins JA, Baer K, Burke S. 2016. Summer 2015 bat survey and radiotelemetry study conducted at the Fort Drum Military Reservation, Jefferson and Lewis counties, New York. Report of Copperhead Environmental Consulting, Inc. to Fort Drum Military Installation, Fort Drum, New York.

Found at DOI: http://dx.doi.org/10.3996/102017-JFWM-080.S4 (9255 KB PDF).

Reference S2. Winhold L, Mann A, Brack V Jr. 2011. Summer mist net surveys for the Indiana bat Myotis sodalis on Fort Drum Military Installation, Jefferson and Lewis counties, New York. Report of Environmental Solutions & Innovations, Inc. to Fort Drum Military Installation, Fort Drum, New York.

Found at DOI: http://dx.doi.org/10.3996/102017-JFWM-080.S5 (684 KB PDF).

This study was supported by the New York State Department of Environmental Conservation and the USFWS. We thank Liliana Dávalos, Kevin Drees, Jeff Foster, Kristjan Mets, Marianne Moore, John Navilio, Katy Parise, Amy Russell, and Colin Sobek for all lab work, preparation, culturing, DNA extraction, testing, and analysis related to P. destructans. We thank John Edwards, Christina Kocer, Robyn Niver, Raymond Rainbolt, and Jonathan Reichard for reviewing early versions of the manuscript. We thank all of the numerous individuals who assisted with data collection. We thank the reviewers and editors at the Journal of Fish and Wildlife Management for assisting in improving our publication.

Any use of trade, firm, or product names is for descriptive purposes only and does not imply endorsement by the U.S. Government.