To test our hypothesis, we studied the invasion dynamics of the WNS fungus by sampling bats of 5 species at 2 hibernacula in central Illinois, USA. We collected samples twice each winter for 2 years (2012–13 and 2013–14). The hibernacula were moderately sized (5–10 hectares, 2–5 m high) abandoned limestone mines that bats use for fall mating and hibernation from September through April. During each visit, we counted all visible bats at each site, which produced complete census data for 4 of the 5 species. Accurately collecting census data for bats of the remaining species (Eptesicus fuscus) was difficult because these bats, unlike those of other species, roosted primarily behind crumbling slabs of rock around mine entrances, which were dangerous and difficult to survey.

During each site visit we sampled 15–20 bats of each species by epidermal swabbing (7). We also sampled the wall or ceiling of hibernacula under, near (10–20 cm), and far from (>2 m) roosting bats by using the same swabbing technique. Samples were tested for P. destructans by using real-time PCR (9); according to a serial dilution experiment, the limit of detection was ≈50 conidia.

Figure 1

Figure 1. Pseudogymnoascus destructans prevalence (±1 SE, calculated from the variance of a binomial distribution sample) over 2 winters, 2012–13 and 2013–14, at 2 sites (diamonds and triangles) in Illinois, USA, on bats...

We obtained 611 samples from bats and 444 from substrate. In early winter of 2012–13, only 1 individual (Myotis septentrionalis) of 129 bats of 5 species sampled was positive for P. destructans, and none of the 46 substrate samples were positive (Figure 1, panels A, C, E). Just 4 months later, in March 2013, prevalence was >85% for bats of 2 species (M. septentrionalis, M. lucifugus), 40%–75% for 2 species (E. fuscus, Myotis sodalis), and 15%–60% for 1 species (Perimyotis subflavus) at the 2 sites (Figure 1, panel A). The prevalence of P. destructans on the substrate under these bats varied from 0% to 67%, and substrate prevalence paralleled fungal prevalence for bats of each species (Figure 1, panel C). Despite widespread apparent infection of bats at this time, none of the 36 substrate samples taken just 10–20 cm from bats were positive for P. destructans (Figure 1, panel E).

In early winter of the next year (late November 2013), patterns differed markedly from those of the previous early winter. P. destructans was already widespread in the environment, found in 70% of samples from under bats, 22% of samples 10–20 cm from bats, and 14% of samples >2 m from bats (Figure 1, panels D, F). Prevalence among bats of 4 species was already >70%, and prevalence among bats of 1 of these species (P. subflavus), for which prevalence at the end of the previous winter had been lowest, was already 85%–100% (Figure 1, panel B). By the end of the second winter, 109 (98%) of 111 bats were positive for P. destructans, and P. destructans was present throughout the hibernacula (in 91% of samples from under bats, 66% of samples near bats, and 44% of samples far from bats) (Figure 1, panels B,D,F).

Figure 2

Figure 2. Complete population counts on a log scale of 4 species of bats at 2 sites in Illinois, USA, over 2 winters, 2012–13 and 2013–14. Diamonds and triangles indicate sites.

Over these 2 years, the effect of WNS on bat populations mirrored the patterns of P. destructans prevalence. During the first winter, declines were limited at the larger site and moderate (50%–75%) at the smaller site (Figure 2). In contrast, over the second winter, counts of M. septentrionalis bats declined by 95%–99% and M. lucifugus bats by 81%–88% (20,000 bats of this species disappeared) (Figure 2, panel A). Populations of bats of the 2 other species also experienced moderate to severe declines in the second year (M. sodalis, 16%–96%; P. subflavus, 47%–73%) (Figure 2, panel B). Declines probably resulted from disease-related deaths because high hibernacula site fidelity makes emigration unlikely (10) and substantial numbers of dead bats were observed at both sites.

Early in the first winter studied, prevalence of P. destructans was very low, and although transmission resulted in most bats harboring P. destructans by winters’ end, declines in bat populations were limited. In contrast, early in the second winter, fungal prevalence among bats was already high and severe communitywide declines occurred over the next 4 months. The earlier timing of exposure in the second year would be expected to increase the effects of WNS because by winter’s end most bats would have been infected and in hibernation for at least 70–100 days (the approximate time between infection and death; 5). Few would be able to survive until spring, when bats cease hibernating and clear the fungus (7).

Patterns of P. destructans distribution in the environment mirrored prevalence among bats and population declines. Early in the first year, when P. destructans was rare on hibernacula substrates, most bats were not infected in early winter, and 4 months later, P. destructans was not detectable in one third of bats of 3 species. However, by the end of the first winter, P. destructans was present on hibernacula substrate under bats, probably resulting from bats shedding P. destructans into the environment. At the beginning of the following winter, P. destructans was widespread in the environment, and nearly all bats had fungus on them. The widespread occurrence of P. destructans in the environment at this time may have contributed to higher prevalence among bats because most bats clear infections during the summer, when their body temperature is too high for P. destructans growth (7,11). Long-term persistence of P. destructans in the absence of bats (8,12) suggests that an environmental reservoir of P. destructans may contribute to WNS persistence, as occurs for other diseases, such as cholera (13).

WNS continues to spread south, west, and north from New York, where it was first detected in 2006, and continues to cause widespread bat population declines. Potential control strategies include development of probiotic treatments (14) and alteration of hibernacula microclimates to make them cooler and drier (3,15). Our results suggest that if P. destructans invasion in other sites is similar to what we documented in Illinois, interventions must be implemented proactively, or quickly after P. destructans invasion, to prevent collapse of bat communities. Reduced bat populations will probably have a negative effect on humans because bats play a useful role in ecosystems by consuming disease vectors and many forest and agricultural insect pests.