2011 SOUTHEASTERN NATURALIST 10(4):609–628 Holocene and Late Pleistocene Bat Fossils (Mammalia: Chiroptera) from Hamilton County, TN, and their Ecological Implications Timothy J. Gaudin1,*, Ashley N. Miller1, Jeremy L. Bramblett1, and Thomas P. Wilson1 Abstract - Chiropteran mandibles from late Pleistocene/Holocene fossil cave localities in Hamilton County were identified in order to examine changes in bat species diversity and population trends over extended periods of time, providing insight into how bats in Southeast Tennessee have responded to major environmental changes over the past 10,000–20,000 years. Generic and species identifications were based on an unpublished key developed by the authors. Measurements of alveolar length (c1–m3) and total length measurements from the symphysis to the condyle were taken for all specimens identifi ed as members of the genus Myotis in an attempt to identify species in this genus. The results of this study failed to confirm those of previous univariate morphological studies, suggesting that multivariate morphometric analyses may be needed to establish a means to differentiate among the species in this genus. Diversity data indicated two patterns of species abundance, with Eptesicus fuscus (Big Brown Bat) dominating some sites and Myotis sp. dominating others. The data suggest, but do not conclusively demonstrate, that a temporal replacement of older Eptesicus faunas by younger, Myotis-dominated faunas has occurred, connected with post-Pleistocene global warming. In addition, a correspondence between human disturbance and bat populations levels was observed. It is very likely that human disturbance has caused bat populations to become extinct in the caves under study, reinforcing the claim of previous researchers that bat population decline is a recent phenomenon that is tightly linked to human disturbance. Introduction The transition between the late Pleistocene Epoch and the early Holocene Epoch is marked by a dramatic change in climate which caused the glacial ice sheets that covered large portions of the continent to gradually retreat (Corgan 1996, Pielou 1991). The warming period at the Pleistocene/Holocene boundary caused major changes in habitat distributions (Pielou 1991). Slowly, the diverse habitats that were common in the Pleistocene were altered as the hypsithermal periods of the early Holocene swept across the continent, with some biogeographical realms shifting northward and others disappearing entirely (Pielou 1991). These ecological changes caused many animal species to change their distribution and many large mammal species to go extinct at this time (Barnosky et al. 2004, Graham et al. 1996, Kurtén and Anderson 1980, Pielou 1991, Schubert et al. 2003). Two major hypotheses exist to explain this extinction pattern: 1) that the climate 1Department of Biological and Environmental Sciences, University of Tennessee at Chattanooga, 615 McCallie Avenue, Chattanooga, TN 37403-2598. *Corresponding author - Timothy-Gaudin@utc.edu. 610 Southeastern Naturalist Vol. 10, No. 4 changes at the end of the Wisconsinan glaciation caused the large mammal extinction, or 2) that prehistoric overkill by humans caused mammal extinctions (Barnosky et al. 2004, Martin 2005, Pielou 1991). Caves play an essential role in our understanding of the Pleistocene Epoch in North America, as they represent the source of the fossils representing the largest percentage of known North American Pleistocene faunas (Kurtén and Anderson 1980, Schubert et al. 2003). They have therefore been critical in allowing paleontologists to document biogeographic and evolutionary patterns in a great variety of organisms (Corgan 1996, Schubert et al. 2003). Cave deposits often preserve Holocene fossils as well, but these are rarely studied by paleontologists because they do not contain extinct taxa (Schubert et al. 2003). However, Holocene sites can provide useful information on the evolution of current ecosystems (Schubert et al. 2003). Tennessee, with its multitude of caves, provides an excellent environment for exploring fossils from caves in both the Pleistocene and Holocene (Corgan 1996). Hamilton County contains a plethora of caves, several of which have been shown to contain Pleistocene and Holocene fossil remains (Bramblett 1998, Corgan 1996, Gaudin and Bramblett 1999, Gaudin et al. 1998, Parmalee 1961). Historically, nearly all of our information concerning the vertebrate fauna of Hamilton County during the Pleistocene and Holocene is based on the fossils of megafaunal and medium-sized vertebrates unearthed in Lookout Mountain Cave by Dr. Henry C. Mercer in 1893 (Corgan 1996, Mercer 1894). Indeed, most of the initial paleontological research on the Pleistocene of North America focused on extinct mammalian megafauna (Kurtén and Anderson 1980). It was not until the 1920s that the number of mammalian microfaunal taxa began to increase substantially (Kurtén and Anderson 1980). Due to the wide biogeographic distribution and broad environmental tolerances of the species found in Mercer’s megafaunal sample (Corgan 1996, Mercer 1894), they are less than ideal indicators of environmental change. Microvertebrates, with their smaller geographic ranges and narrower ecological niches, could provide a better understanding of how vertebrate communities responded to environmental changes that occurred during the Pleistocene and Holocene. Several years ago microfaunal paleontological fieldwork was initiated in caves at Lookout Mountain, TN (Bramblett 1998; Bramblett and Gaudin 2001; Gaudin and Bramblett 1999; Gaudin et al. 1998, 1999) and in nearby parts of Hamilton County (Ooltewah High School, Ooltewah, TN), and a large and diverse sample of Pleistocene microvertebrate remains were recovered. Included in these faunas were over 10,000 skeletal elements from bats, including partial skulls, mandibles, and isolated postcranial remains (Bramblett and Gaudin 2001, Gaudin and Bramblett 1999). A sample of these bat fossils were identified (Bramblett and Gaudin 2001, Gaudin and Bramblett 1999), demonstrating the presence of at least six extant species of bats, most of which are cave-dwelling for at least part of the year: Myotis septentrionalis (Trouessart) (Northern Long-eared Bat), M. grisescens A.H. Howell (Gray Bat), Perimyotis subflavus 2011 T.J. Gaudin, A.N. Miller, J.L. Bramblett, and T.P. Wilson 611 (F. Cuvier) (Tricolored Bat), Corynorhinus rafinesquii Lesson (Rafinesque’s Big-eared Bat), Eptesicus fuscus (Beauvois) (Big Brown Bat), and Lasiurus sp. (L. borealis (Müller) [Eastern Red Bat] or L. seminolus (Rhoads) [Seminole Bat]) (Bramblett and Gaudin 2001, Gaudin and Bramblett 1999). Among these species, C. rafinesquii is currently uncommon and of special concern in parts of its range, and M. grisescens is a federally listed endangered species (Choate et al. 1994, Harvey et al. 1999). Although a small population of P. subflavus persists in Ooltewah High School Cave, we found no evidence that bat colonies are currently dwelling in any of the Lookout Mountain caves in which fossils were discovered (Bramblett 1998). The presence of extensive bat colonies in these caves in the Pleistocene and Holocene, and their absence today, raises intriguing questions regarding the causes of their disappearance. Obvious human disturbance is evident at all our Lookout Mountain Cave localities. These disturbances range from a railroad track that runs through one entrance to guided tours at another entrance. Human perturbation is thought to be a major contributor to declines in cave bat populations by disturbing hibernation or nursery sites (Tuttle 1979). Tuttle (1979) used the size and color of stains from bat guano (or other bodily secretions) on the ceilings of caves as an indicator of the size of historical bat populations roosting in the caves. The estimates of current population size were obtained by taking the number of square meters covered by new guano and multiplying it by the mean clustering density of 1828/m2 (Tuttle 1975, 1976, 1979). Historic colony size was calculated from the area covered by old guano deposits or the area of staining on the cave roof (Tuttle 1979). Tuttle’s work suggested that M. grisescens, a species of bat found among the fossil bats from the Lookout Mountain caves, had recently experienced a dramatic population decline (Tuttle 1979). We note that this inference depends on the assumption of roost fidelity, i.e., that colonies roost in the same part of the cave ceiling each year (Tuttle 1979). Though Tuttle may have observed roost fidelity during the six years of his study (Tuttle 1979), it is possible that over centuries or millennia the bats slowly migrated across the cave ceiling, using different parts as roosting areas. If the latter were true, Tuttle’s (1979) inferences about the long-term population trends in this species might need to be reconsidered. Paleontology may offer an alternative method for examining long-term population trends by providing data over a longer time frame than guano stains will allow. Furthermore, paleontological studies offer an independent source of data via the preservation of actual remains. Because most bat species from the Pleistocene and Holocene of North America are still extant (Kurtén and Anderson 1980), comparison of information on the known taxonomic diversity and ecological characteristics of the modern bat fauna (Barbour and Davis 1969; Best and Jennings 1997; Caceres and Barclay 2000; Choate et al. 1994; Decher and Choate 1995; Fujita and Kunz 1984; Harvey 1992; Harvey et al. 1999; Jones 1977; Kunz and Martin 1982; Kurta and Baker 1990; Shump and Shump 1982a, b) to paleontological cave assemblages has the potential to yield 612 Southeastern Naturalist Vol. 10, No. 4 insight into a number of interesting questions. For example, such comparisons can give us insight into how changes in biogeographic distributions and population levels over time are correlated with environmental changes. The goal of the present study is to use paleontology as a tool to contribute to our knowledge of the historical ecology of bats. We identified chiropteran mandibles from known fossil localities in order to examine changes in population and species diversity over extended periods of time. This information in turn should provide insight into how bats in Southeast Tennessee have responded to major environmental changes, including climate change, habitat change, and human disturbance over the past 10,000–20,000 years. Methods Chiropteran mandibles were identified based on a taxonomic key for lower jaws of Holocene bats from the southeastern United States prepared by three of us (T.J. Gaudin, J.L. Bramblett, and A.N. Miller) (see Supplemental Appendix 1, available online at https://www.eaglehill.us/SENAonline/suppl-files/s10-4 -882-Gaudin-s1, and, for BioOne subscribers, at http://dx.doi.org/10.1656/ S882.s1). This key allowed for identifications of most specimens to the species level. However, only one species (Myotis septentrionalis) within the genus Myotis was distinguishable using meristic characteristics. A quantitative method for determining species identifications within the genus Myotis was used by Guilday et al. (1977) to identify specimens belonging to the species Myotis grisescens and Myotis leibii (Audubon and Bachman) (Eastern Small-footed Bat). Their method involved measuring alveolar length between alveolus of the lower canine and the last lower molar (c1 to m3). These measurements were plotted on a histogram, revealing distinct peaks associated with the smallest (M. leibii) and the largest (M. grisescens) species of Myotis (Guilday et al. 1977). We repeated Guilday et al.’s (1977) measurements for all specimens identified as members of the genus Myotis. Total mandibular lengths from the symphysis to the condyle were also taken on all Myotis specimens for purposes of comparison. All measurements were obtained using Mitutoyo® dial calipers, and recorded to the nearest 0.01 mm. The measurement data were then used to construct histograms using Microsoft Excel®. No current method can be used to unambiguously identify skeletal remains of the other southeastern US species of Myotis to the species level (M. sodalis Miller and Allen [Indiana Bat], M. austroriparius (Rhoads) [Southeastern Bat], and M. lucifugus (LeConte) [Little Brown Bat]). Therefore, we were unable to distinguish these species in this study. Taxonomic identifications and provenance of over 900 unidentified fossil bat mandibles from four localities were recorded in a database at the University of Tennessee at Chattanooga Natural History Museum. Two of the localities lie in separate portions of Lookout Mountain Cave (herein designated as RR, RF) (Barr 1961, Corgan 1996). Another unnamed locality was designated Cave Without Name (CWN) by Gaudin et al. (1998, 1999). All of these localities lie at the north 2011 T.J. Gaudin, A.N. Miller, J.L. Bramblett, and T.P. Wilson 613 end of Lookout Mountain near Chattanooga, TN. The final locality is Ooltewah High School Cave (OHS) in Ooltewah, TN (Table 1). After all bat fossils were identified and cataloged, the diversity at specific recovery sites within each cave locality was determined using pie charts created in Microsoft Excel®. Minimum number of individuals (MNI) (based on right mandibles) were initially calculated in order to estimate population sizes for the various bat species at each site (Benton et al. 1994). However, total number of elements ultimately was used because MNI yielded sample sizes that were too small for analysis. The total number of elements for each species at a given site was then used to calculate heterogeneity and evenness measures from each site using Simpson’s measure of evenness (Simpson 1949), Camargo’s index of evenness (Carmargo 1993), and Smith and Wilson’s index of evenness (Smith and Wilson 1996) using the computer program Methods® (Krebs 2000). The Methods® program was also used to generate Euclidean distances, a Bray-Curtis metric (Bray and Curtis 1957), a Canberra metric (Lance and Williams 1967), and percent similarities (Renkonen 1938) for each locality except OHS cave. These latter tests require data from at least three sites. Unfortunately, OHS cave did not have enough sites to run these analyses. Evenness measurements should indicate the degree to which relative abundances of individuals among the different species are comparable (Krebs 2000). Euclidean distance, Bray-Curtis metric, Canberra metric, and percent similarity give information about the degree of relative similarity among the sites at a given locality with regards to species composition. Several sites were omitted from analysis for the Euclidean distance, Bray-Curtis metric, Canberra metric, and percent similarity measures from both the RF (site 5) and RR (sites 7 and 8) localities. The small number of species recovered from these sites interfered with the function of the pertinent equations. Table 1. Cave locality information. Sample Cave Elevation Latitude and longitude Site Site age (YPB) Sediment type size (n) CWN 660’ 35.024°N, 85.342°W 1 14,459 ± 786 Yellow sand 26 2 14,459 ± 786 Yellow sand 13 4 14,459 ± 786 Yellow sand 32 RF 660’ 35.021°N, 85.338°W 1 14,811 ± 682 Orange clay 25 2 14,811 ± 682 Orange clay 9 4 14,811 ± 682 Orange clay 6 5 14,811 ± 682 Orange clay 1 6 25,458 ± 2100 Orange clay 21 RR 700’ 35.021°N, 85.338°W 1 16,148 ± 483 Orange clay 614 2 Unknown Dark silty clay 86 3 >10,000 Orange clay 20 4 Unknown Dark silty clay 12 5 less than 500 Orange clay 18 7 Unknown Dark silty clay 1 8 Unknown Dark silty clay 4 OHS 800’ 35.095°N, 85.066°W 1 less than 10,000 Brown sandy silt 4 614 Southeastern Naturalist Vol. 10, No. 4 Ages of sites within each locality were determined by radiocarbon dating and faunal analysis. AMS Radiocarbon dating was performed by Rafter Radiocarbon Laboratory (Lower Hutt, New Zealand). Elements used for the radiocarbon analysis were selected from Pleistocene species, where possible, in an attempt to approximate the maximum ages for each site. Dates were obtained for three sites in RR: RR1 at ≈16,000 ybp, based on radiocarbon dating (16,148 ± 483 ybp, date based on dermal scute from the extinct Dasypus bellus (Simpson) [Beautiful Armadillo], sample ID number R 24720/1); RR3 at >10,000 ybp, based on the presence of D. bellus (Kurtén and Anderson 1980); and RR5 at less than 500 ybp, based on the presence of Mus musculus L. (House Mouse). The site at the OHS cave locality was dated at less than 10,000 ybp based on the presence of Blarina carolinensis (Bachman) (Southern Short-tailed Shrew), which, according to Klippel and Parmalee (1982), likely arrived in middle Tennessee in the Holocene. Radiometric dates were obtained for two sites in RF: RF1 at ≈15,000 ybp (14,811 ± 682 ybp, date obtained from molar of extinct peccary Mylohyus sp., R 24888/2); and RF6 at ≈25,500 ybp (25,458 ± 2100 ybp, date obtained from radius of Marmota monax L. (Woodchuck), R 24888/3). However, sites RF1–5 were part of the same streambed, and fossils at all five localities are recovered only in the upper few centimeters of sediment. We therefore believe that all five sites can be treated as effectively contemporaneous. Similarly, a radiometric date was obtained for only one site in CWN (CWN1 dated at 14,459 ± 786 ybp, date obtained from upper molar of Mylohyus sp., R 24888/1), but because CWN2 and CWN4 were part of the same streambed, with fossils found only in the upper few centimeters of sediment, they too were treated as contemporaneous. Note the deposits from the RR (16,000 ybp to less than 500 ybp) and RF (25,500 ybp to 15,000) localities likely exhibit a relatively small temporal overlap, and that both encompass a broad range in time, from 10 to 15 thousand years (Table 1). Results A total of 909 bat mandibles were identified from the four cave localities. The most common genus within the sample was Myotis, with 439 elements. The most common species identified was Eptesicus fuscus, with 268 elements. Other taxa that occurred in smaller numbers included species within the genus Corynorhinus (68 elements: C. sp., 42 elements; cf. C. sp., 20 elements; C. rafinesquii, 6 elements), Perimyotis subflavus (29 elements), and Lasiurus borealis or L. seminolus (6 elements; cf. L. borealis or L. seminolus, 1 element). A portion of the sample was unidentifiable at the genus or species level (97 elements). Of those 439 specimens belonging to the genus Myotis, 309 elements were sufficiently preserved to obtain mandibular measurements. Total mandibular lengths (symphysis to condyle) for this sample ranged from 9.38 mm to 11.60 mm (Fig. 1). Alveolar length (c1 to m3) ranged from 4.72 mm to 7.50 mm (Fig. 2). The histograms created based on these measurements show several distinct peaks (Figs. 1, 2). For total mandibular length, six peaks are noted at 9.4 2011 T.J. Gaudin, A.N. Miller, J.L. Bramblett, and T.P. Wilson 615 mm, 9.7– 9.8 mm, 10.1 mm, 10.4 mm, 11.1 mm, and 11.4 mm (Fig. 1). A total of five peaks are present among the alveolar length data. These occur at 5.0 mm, 5.5 mm, 5.7–5.8 mm, 6.2 mm, and 6.5 mm respectively (Fig. 2). Because Myotis septentrionalis was the only species in this genus that could be identified with a meristic (i.e., size-independent) characteristic, specieslevel variation in total mandibular and alveolar length measures was examined separately. M. septentrionalis exhibited broad variation in both measurements, overlapping much of the range of the total data sample, with mandibular lengths ranging from 10.3–11.4 mm (mode of 10.9 mm), and alveolar lengths (c1–m3) ranging from 5.5–6.9 mm (mode of 6.5 mm). When M. septentrionalis was removed from the total sample (Figs. 3, 4), the peak at 11.4 mm in the total mandibular length data was eliminated, whereas the other five original peaks remained intact (Fig. 3). For the alveolar length data, removal of M. septentrionalis left all peaks intact, but resulted in a decrease in the magnitude of the peaks at 5.5 mm, 6.2 mm, and 6.5 mm (Fig. 4). The total number of individuals, site diversity, and site age are summarized in Tables 1 and 2. Charts illustrating the abundance of species recovered at each Figure 1. Histogram showing distribution of mandibular lengths (measured from symphysis to the condyle) among all Myotis specimens measured in the present study (n = 120). Figure 2. Histogram showing distribution of alveolar lengths (measured from c1 to m3) among all Myotis specimens measured in the present study (n = 309). 616 Southeastern Naturalist Vol. 10, No. 4 site (Figs. 5–8) appear to fall into two categories. Half the sites are dominated by Eptesicus fuscus, with E. fuscus comprising at least 50% of the total number of recovered specimens (Figs. 5, 6). These E. fuscus-dominated sites include all the sites at the Cave without name (CWN) locality and the Ruby Falls entrance to Lookout Mountain Cave (RF) locality. The remaining bat faunas are dominated by members of the genus Myotis (Figs. 7–9). Faunas where Myotis comprised Table 2. Summary of similarity metrics at each locality. BC = Bray-Curtis metric, CM = Canberra metric, ED = Euclidean distance, and PS = percent similarity. Locality Sites BC CM ED PS CWN 1 vs. 2 0.46 0.15 4.55 69.2 1 vs. 4 0.86 0.91 1.93 95.0 2 vs. 4 0.40 0.21 6.31 69.2 RF 1 vs. 2 0.51 0.31 6.81 80.0 1 vs. 4 0.33 0.42 7.35 92.0 1 vs. 6 0.83 0.57 2.45 76.8 2 vs. 4 0.67 0.57 1.50 80.0 2 vs. 6 0.45 0.49 7.09 64.8 4 vs. 6 0.38 0.83 8.00 84.8 RR 1 5 1 vs. 2 0.23 0.21 81.83 68.3 1 vs. 3 0.06 0.13 94.82 77.3 1 vs. 4 0.04 0.13 96.59 59.8 1 vs. 5 0.06 0.14 95.30 57.0 2 vs. 3 0.38 0.43 14.93 79.4 2 vs. 4 0.24 0.41 16.60 64.0 2 vs. 5 0.35 0.46 15.09 73.6 3 vs. 4 0.56 0.63 2.41 65.0 3 vs. 5 0.68 0.61 2.10 65.0 4 vs. 5 0.80 0.94 1.61 86.1 Figure 3. Histogram showing distribution of mandibular lengths (measured from symphysis to the condyle) among all Myotis specimens, except M.septentrionalis, measured in the present study (n = 101). 2011 T.J. Gaudin, A.N. Miller, J.L. Bramblett, and T.P. Wilson 617 at least 50% of recovered specimens are found at all the sites from the railroad entrance to Lookout Mountain Cave (RR) and the Ooltewah High School Cave locality (OHS). Among all the sites excavated, RR1 had the largest number of Figure 4. Histogram showing distribution of alveolar lengths (measured from c1 to m3) among all Myotis specimens, except M. septentrionalis, measured in the present study (n = 258). Figure 5. Species diversity based on total number of individuals recovered per species at each of the various sites within the Cave without Name (CWN) locality: a. Site 1, b. Site 2, and c. Site 4. 618 Southeastern Naturalist Vol. 10, No. 4 Figure 7 (opposite page). Species diversity based on total number of individuals recovered per species at each of the various sites within the Lookout Mountain Cave, Railroad entrance (RR) locality; a. Site 1 and b. Site 2. Figure 6. Species diversity based on total number of individuals recovered per species at each of the various sites within the Lookout Mountain Cave, Ruby Falls (RF) locality: a. Site 1, b. Site 2, c. Site 4, and d. Site 6. individuals, the highest level of species diversity, and included all species of Chiroptera found within the total sample (Fig. 7a, Table 1). Information on the similarity among sites (excluding OHS) obtained using Euclidean distance (ED), Bray-Curtis metric (BC), Canberra metric (CM) and percent similarity (PS) indicate varying degrees of similarity (Table 2). Among the sites within CWN, sites 1 and 4 appear to be the most similar, with a 95% similarity. Sites 1 and 4 also have BC and CM values that approach 1, and a markedly smaller ED than other site-by-site comparisons. This similarity can also be observed by comparing the species list for these sites (Fig. 5). Strong PS are seen throughout all sites in RF, with the lowest being ≈65% for site RF2 versus RF6. Site RF6 exhibits a stronger PS value when compared with site RF4 (≈85%) than it does when compared to site RF2, which is similar in age to site RF4 (Table 1). The ED values for site RF6 are relatively high when compared with sites RF2 and RF4 (7.09 and 8.00, respectively), but markedly lower 2011 T.J. Gaudin, A.N. Miller, J.L. Bramblett, and T.P. Wilson 619 620 Southeastern Naturalist Vol. 10, No. 4 when compared with site RF1 (2.45). Some inconsistencies can be noted when comparing results across indices. CM seemed to indicate that site RF4 and RF6 are the most similar. Values from BC showed sites RF1 and RF6 as most similar, whereas PS indicated sites RF1 and RF4 are most similar. Sites RF2 and RF4 had the smallest ED values. Figure 8. Species diversity based on total number of individuals recovered per species at each of the various sites within the Lookout Mountain Cave, Railroad entrance (RR) locality: a. Site 3, b. Site 4, c. Site 5, and d. Site 8. Figure 9. Species diversity based on total number of individuals recovered per species at each of the various sites within the Ooltewah High School Cave (OHS) locality. 2011 T.J. Gaudin, A.N. Miller, J.L. Bramblett, and T.P. Wilson 621 The railroad entrance to Lookout Mountain sites also exhibited fairly high PS values (57% or greater). ED, CM, and BE indicate similarity among all sites except site RR1, although PS values for site RR1 are well within the range of the other site-by-site comparisons. The most similar sites appear to be RR4 and RR5, with the smallest ED (1.61), BC and CM values very close to 1, and a very high PS (86.1%). This observation is reinforced by a comparison of species abundance patterns for those sites (Fig. 8b, c). Information on the heterogeneity and evenness within each locality was obtained using Smith and Wilson’s index (SW), Carmargo’s index (C), and Simpson’s measure of evenness (S) (Table 3). These indices indicated that most localities had a moderate level of biodiversity. High levels of biodiversity were noted using SW, C, and S for RR1 (SW = 0.204, C = 0.379, S = 0.390) and RR2 (SW = 0.316, C = 0.292, S = 0.241). Results from SW also indicated that RF6 (SW = 0.267, C = 0.548, S = 0.55) had high levels of biodiversity, whereas the other indices indicated average levels of biodiversity at this site. Relatively low biodiversity was evident using SW, C, and S for RF4 (SW = 0.924, C = 0.833, S = 0.9) and OHS1 (SW = 0.813, C = 0.75, S = 0.8). The value of 1.0 for RR8 using SW, C, and S reflected the fact that sample size for all species were equal at this site. Discussion Although, as noted previously, only Myotis septentrionalis could be distinguished from other Myotis species based on meristic characters, we were able to recognize some Myotis grisescens and M. leibii specimens based on morphological measurements. However, the size distribution within our sample differs dramatically from that of Guilday et al. (1977). Guilday et al. Table 3. Summary of heterogeneity and evenness at each site. S = Simpson's measure of evenness, C = Carmago's index, and SW = Smith and Wilson's index. Sites S C SW CWN1 0.691 0.590 0.569 CWN2 0.494 0.577 0.696 CWN4 0.603 0.563 0.341 RF1 0.590 0.573 0.566 RF4 0.900 0.833 0.924 RF5 0.667 0.667 0.743 RF6 0.550 0.548 0.267 RR1 0.390 0.379 0.204 RR2 0.241 0.292 0.316 RR3 0.476 0.500 0.515 RR4 0.600 0.600 0.680 RR5 0.476 0.511 0.576 RR8 1.000 1.000 1.000 OHS1 0.800 0.750 0.813 622 Southeastern Naturalist Vol. 10, No. 4 (1977) identified any specimen with an alveolar length of 5.5 mm or less as pertaining to M. leibii and any specimen with an alveolar length of 6.6 mm or greater as belonging to M. grisescens (Guilday et al. 1977). In contrast, in our study, the lower alveolar peak occurred at 5.0 mm, and no clear upper alveolar length peak was present, except for a peak at 6.5 mm (Fig. 2) that fell within the range of M. septentrionalis (5.5–6.9 mm) and therefore could not be attributed to M. grisescens. When the total length data were examined, a clear upper peak could not be seen once M. septentrionalis was removed from the histogram (Fig. 4). Nevertheless, we were able to assign specimens above the range for total mandibular length in M. septentrionalis (11.5 mm or greater) to M. grisescens. Specimens occurring at or below the lower peaks in both total mandibular length (9.5 mm or less) and alveolar length (5.0 mm or less) were assigned to M. leibii. Although differences in the age or the mode of formation of the deposits may exist between the present study and that of Guilday et al. (1977), it is not readily apparent how these differences would translate into differences in the distribution of measurements found in both studies. The range in locality age (<500 ybp to 25,500 ybp) is much greater in our study than in that of Guilday et al. (1977). Guilday et al. (1977) estimate that the Clark’s Cave deposit is somewhere between 20,000 and 11,000 ybp. Thus, their fauna may be older than some of our remains. Therefore, it is conceivable that the relevant species of Myotis have changed size over time, as is known to occur in other mammalian lineages, which often become smaller in the warming post-Pleistocene climate (Kurtén and Anderson 1980). This hypothesis would explain the apparent reduction in the size of M. leibii mandibles, but would not account for the lack of a clear upper peak for M. grisescens in our analysis. It is also possible that both analyses have erred in assuming that Holocene species represent a reasonable proxy for Pleistocene diversity. It is certainly possible that different Myotis species, either extinct varieties or Recent species with a different modern geographic distribution, inhabited these areas at various times in the Pleistocene, skewing the measurement distribution of one or both analyses. Whatever the cause, taken together, the results of the two studies imply that simple length measures may not be sufficient even for identification of the largest and the smallest Myotis species. Moreover, neither we nor Guilday et al. (1977) were able to distinguish between the medium-size Myotis species from the southeast [M. austroriparius, M. sodalis, and M. lucifugus]. It seems clear that multivariate morphological studies are needed to establish a means to differentiate among the species in this genus. Such multivariate analyses are beyond the scope of this study. Community-level numeric analysis seems to suggest close similarity among sites within CWN (Table 2). This similarity makes sense considering that sites in CWN occur within a single dry streambed. The fact that sites CWN1 and CWN4 are physically the two most distant sites at the locality but are ecologically the most similar seems to confirm that all sites at this locality are uniform in age and composition, and may once have been part of a 2011 T.J. Gaudin, A.N. Miller, J.L. Bramblett, and T.P. Wilson 623 larger metapopulation. Interestingly, similar intra-site homogeneity is evident for RF, even though there appears to be a 10,000-year age difference between RF1–4 and RF6. Some RF sites appear to be more similar to RF6 than other RF sites of similar age, and all sites are dominated by Eptesicus fuscus. Therefore, based on this sample, site age does not appear to have a strong effect on species composition for RF sites. The RR locality also exhibits substantial similarity among sites for some, but not all, measurements. Though species composition among sites was similar (Figs. 7 and 8), ED, CM, and BC indicated a sizable difference between site 1 and the other sites (Table 2). This difference may be due to the large sample size from site 1 (614 elements) when compared to other sites (4–86 elements). The large sample sizes of RR1 and RR2 may also help explain the higher level of biodiversity noted for both localities using SW, C, and S. Similarly, the relatively small sample size (4–6 elements) of RF4, RR8, and OHS1 likely explains the low levels of diversity seen using SW, C, and S for these localities. The variation in the abundance of particular bat genera and species among localities is apparent from the species distributions pie charts shown in Figures 5–9. These figures indicate that E. fuscus dominates sites within CWN and RF, whereas Myotis sp. dominates sites at RR and OHS. The reason for this difference in abundance is not clear. RF and RR are both part of the same cave system and share similar external environments, indicating that external environments probably do not play a role in the pattern of dominance observed. Cave size also does not seem to be a factor in determining species abundance. RF and RR are part of a single large cave system, but have distinctly different species abundance patterns. OHS and CWN are relatively small caves, but are also dominated by different species. Habitat preference of extant bat species might help explain species distributions within our sample. Some bat species, like Myotis grisescens, roost only in caves (Barbour and Davis 1969, Choate et al. 1994, Harvey 1992, Harvey et al. 1999). Other species like M. leibii, E. fuscus, and Perimyotis subflavus, occur in a wider range of habitats that include abandoned mines, man-made structures, rock crevices, trees, etc. (Barbour and Davis 1969, Choate et al. 1994, Harvey 1992, Harvey et al. 1999). Although one species, L. seminolus, will occasionally roost in caves, and individuals of several other species of southeastern Lasiurus are sometimes found in and around cave entrances, the preferred roosting habitat for most southeastern Lasiurus are trees (Barbour and Davis 1969, Choate et al. 1994, Harvey 1992, Harvey et al. 1999), which almost certainly explains why Lasiurus sp. represents only a small portion of the sample. Interestingly, exclusively cave-dwelling species do not make up a majority of our sample. However, with the exception of Lasiurus, there is no clear correlation between habitat preferences and the pattern of species abundance in our sample. In fact, Myotis sp. and E. fuscus were found to dominate cave localities that appear to conflict with their general habitat preferences. Many species of Myotis prefer cool but not freezing winter hibernacula, and tend to be more abundant in 624 Southeastern Naturalist Vol. 10, No. 4 deeper portions of the caves they inhabit (Barbour and Davis 1969, Best and Jennings 1997, Choate et al. 1994, Decher and Choate 1995, Harvey 1992, Harvey et al. 1999)—M. leibii is an exception to this generalization, but this rare, cold-tolerant species apparently represents only a small portion of our sample. On the other hand, E. fuscus is very cold tolerant and prefers cooler, even freezing hibernacula (Barbour and Davis 1969, Choate et al. 1994, Harvey 1992, Harvey et al. 1999). It is therefore often found near the entrances of caves. Yet in our own sample, Myotis were prevalent from the front of the Lookout Mountain cave system (RR locality), whereas E. fuscus were abundant at the back of the same cave system (RF locality). Colony size may also contribute to the observed pattern in species abundances. Several bat species present in our sample are characterized by large colony sizes of 100 to over 1000 individuals. This pattern is typical for M. grisescens, M. lucifugus, M. sodalis, and E. fuscus (Barbour and Davis 1969, Choate et al. 1994, Decher and Choate 1995, Harvey 1992, Harvey et al. 1999, Kurta and Baker 1990). It is perhaps not surprising then that these two genera, Myotis and Eptesicus, dominate the overall sample, although it does not explain why particular faunas are dominated by one genus or the other. Our remains also incorporate species (Corynorhinus sp., Lasiurus sp., Perimyotis subflavus, M. septentrionalis, and M. leibii) that roost singly or in small colonies less than 100 individuals (Barbour and Davis 1969; Best and Jennings 1997; Caceres and Barclay 2000; Choate et al. 1994; Fujita and Kunz 1984; Harvey 1992; Harvey et al. 1999; Kunz and Martin 1982; Shump and Shump 1982a, b). One might expect these taxa to dominate the small samples obtained from smaller caves because small caves cannot accommodate large colonies. However, this was not the case in our study. One small cave system, CWN, was dominated by E. fuscus remains, whereas the other small cave system, OHS, yielded mostly remains from Myotis sp., although none of the OHS Myotis specimens could be identified to species. M. septentrionalis represents a sizable portion of the sample from the Lookout Mountain cave system (RR and RF), yet it has colonies of no more than 100 individuals (Barbour and Davis 1969, Caceres and Barclay 2000, Choate et al. 1994, Harvey 1992, Harvey et al. 1999). The prevalence of M. septentrionalis may be an indication that several of these cave sites preserve a mixed-age assemblage. The large sample of M. septentrionalis elements is likely the result of the accumulation of elements over a long period of time. Although the sites themselves do not provide detailed stratigraphic information, time averaging of M. septentrionalis specimens could be demonstrated through AMS radiocarbon dating of multiple specimens, but this is expensive and is beyond the scope of this study. It is possible that the differential pattern of species abundance at different sites can be explained as the result of temporal changes associated with overall climate change in the area. Most of the E. fuscus-dominated localities (CWN1, 2, and 4 and RF1, 2, 4, and 6) are similar in age, having been dated 2011 T.J. Gaudin, A.N. Miller, J.L. Bramblett, and T.P. Wilson 625 to an age of about 14,000 ybp (14,811 ± 682 ybp [RF1, 2, and 4] and 14,459 ± 786 ybp [CWN sites]). However, site RF6, which is also E. fuscus dominated, is ≈10,000 years older (25,458 ± 2100 ybp) than the other sites. The age range for Myotis-dominated sites (RR and OSH) is even more substantial. Within two RR localities alone, there are sites radiocarbon dated to >16,000 ybp (16, 148 ± 483 ybp [RR1]) and sites dated faunally to <500 ybp (RR5). The OHS cave was faunally dated to <10,000 ybp. The Eptesicus-dominated faunas are by and large older than the Myotis-dominated faunas. Moreover, the oldest site, RF6, has the highest percentage representation of E. fuscus of any of our sites (20/21 specimens), although the sample size is modest. The youngest site (RR5) has the highest proportion of Myotis remains (14/18 specimens), and the other post- Pleistocene site (OHS1) is 75% Myotis (3/4 specimens), although the sample sizes at both sites are small. In the Pleistocene sites from the RR locality, Myotis comprises closer to 50% of the respective faunas (e.g., 11/20 specimens at RR3, 303/614 at RR1). It is tempting to interpret this pattern as a temporal replacement of E. fuscus by species in the genus Myotis. Furthermore, this temporal change would be consistent with changes in climate that are occurring simultaneously. The substantial climatic warming that has occurred since the mid-Pleistocene (Pielou 1991) could be construed as favoring the less-coldtolerant species of Myotis over the more-cold-tolerant E. fuscus. However, it is important to note that the Myotis-dominated faunas overlap the E. fuscusdominated faunas by at least two thousand years, and that only four sites in the study have been carbon dated. Therefore, at this point, we cannot conclude with certainty that the patterns of species abundance are created by temporal changes in climate. While global warming might explain faunal changes in the distant past, and may be inferred as a continuing source of faunal change in the near future, it has been suggested that recent changes in bat species diversity and abundance are attributable to human-related disturbances. The results of our study would seem to reinforce this conclusion. Our fossils provide evidence that the Lookout Mountain cave system supported large bat populations for over 25,000 years. Bat remains are recovered from deposits formed as recently as <500 ybp (RR5), indicating that the disappearance of bats from this system is a very recent phenomenon. It is known that as human population levels rapidly increased in the area over the past 500 years, human disturbance of the caves has increased. In the Lookout Mountain cave system, these disturbances include saltpeter mining, the building of a railroad tunnel through the RR locality, and even the use of the RF locality as a tourist site. These data therefore strongly implicate anthropogenic disturbance as a major source of bat disappearance in the Lookout Mountain cave system. Thus, these results reinforce Tuttle’s (1979) claim, based on different lines of evidence (guano stains), that bat population decline in southeast Tennessee is a recent phenomenon that is tightly linked to human disturbance. It also reemphasizes the need to decrease human disturbance in order to prevent further declines in extant bat abundance and diversity. 626 Southeastern Naturalist Vol. 10, No. 4 Acknowledgments We wish to thank the CSX Corporation for facilitating access to several of the cave sites where these bat specimens were collected. We thank P. Millener, K. Ballew, P. van Alstyne, W. Stevens, M. Stevens, M. Konertz, J. England, and many of T. Gaudin’s former students for assistance in collecting and processing specimens. AMS radiocarbon dating of specimens from several cave sites was performed by Rafter Radiocarbon Laboratory (Lower Hutt, New Zealand), and we thank Dr. Phil Millener for his assistance in obtaining these dates. For access to specimens used in creating our taxonomic key and identifying our bat specimens, we thank W. Klippel, University of Tennessee at Knoxville, M. Kennedy of the University of Memphis, D. Ekkens of Southern Adventist University, and E. McGhee of the University of Georgia. For their comments on a previous draft of this manuscript, we thank E. Carver and M. Santiago. A. Miller’s research was supported by a Provost Student Research Award from the University of Tennessee at Chattanooga. Literature Cited Barbour, R.W., and W.H. Davis. 1969. Bats of America. The University Press of Kentucky, Lexington, KY. 286 pp. Barnosky, A.D., P.L. Koch, R.S. Feranec, S.L. Wing, and A.B. Shabel. 2004. Assessing the causes of late Pleistocene extinctions on the continents. Science 306:70–75. Barr, T.C. 1961. Caves of Tennessee. Tennessee Division of Geology, Bulletin 64, Nashville, TN. 567 pp. Benton, M.J., G. Warrington, A.J. Newell, and P.S. Spence. 1994. A review of the British middle Triassic tetrapod assemblages. Pp. 131–160, In N.C. Fraser and H.-D. Sues (Eds.). In the Shadow of Dinosaurs: Early Mesozoic Tetrapods. Cambridge University Press, Cambridge, UK. 435 pp. Best, T.L., and J.B. Jennings. 1997. Myotis leibii. Mammalian Species 547:1–6. Bramblett, J.L. 1998. A survey of Pleistocene fossil and Holocene subfossil vertebrates from caves on Lookout Mountain. Departmental Honors Thesis. University of Tennessee Chattanooga. Chattanooga, TN. 44 pp. Bramblett, J.L., and T.J. Gaudin. 2001. Late Pleistocene bats from Hamilton County, Tennessee. Abstract, 11th Colloquium on Conservation of Mammals in the Southeastern United States. Memphis, TN. Bray, J.R., and J.T. Curtis. 1957. An ordination of the upland forest communities of southern Wisconsin. Ecological Monographs 27:325–349. Caceres, M.B., and M.R. Barclay. 2000. Myotis septentrionalis. Mammalian Species 634:1–4. Carmago, J.A. 1993. Must dominance increase with the number of subordinate species in competitive interactions? Journal of Theoretical Biology 161:537–542. Choate, J.R., J.K. Jones, Jr., and C. Jones. 1994. Handbook of Mammals of the Southcentral States. Louisiana State University Press, Baton Rouge, LA. 304 pp. Corgan, J.X. 1996. Tennessee’s Prehistoric Vertebrates. Tennessee Division of Geology, Bulletin 84, Nashville, TN. 170 pp. Decher, J., and J.R. Choate. 1995. Myotis grisescens. Mammalian Species 510:1–7. Fujita, M.S., and T.H. Kunz. 1984. Pipistrellus subflavus. Mammalian Species 228:1–6. Gaudin, T.J., and J.L. Bramblett. 1999. Preliminary report on late Pleistocene bats from Lookout Mountain, Tennessee. ASB Bulletin 46:167. 2011 T.J. Gaudin, A.N. Miller, J.L. Bramblett, and T.P. Wilson 627 Gaudin, T.J., J.L. Bramblett, P.R. Millener, P.C. Van Alstyne, and K. Ballew. 1998. New late Pleistocene and Holocene microvertebrate localities from Lookout Mountain, Tennessee. Journal of Vertebrate Paleontology 18:45A. Gaudin, T.J., J.L. Bramblett, P.R. Millener, P.C. Van Alstyne, and K. Ballew. 1999. New late Pleistocene and Holocene microvertebrate localities from Lookout Mountain, Tennessee. Journal of the Tennessee Academy of Science 74:26. Graham, R.W., E.L. Lundelius, Jr., M.A. Graham, E.K. Schroeder, R.S. Toomey III, E. Anderson, A.D. Barnosky, J.A. Burns, C.S. Churcher, D.K. Grayson, R.D. Guthrie, C.R. Harington, G.T. Jefferson, L.D. Martin, H.G. McDonald, R.E. Morlan, H.A. Semken, Jr., S.D. Webb, L. Werdelin, and M.C. Wilson. 1996. Spatial response of mammals to late Quaternary environmental fluctuations. Science 272:1601–1606. Guilday, J.E., P.W. Parmalee, and H.W. Hamilton. 1977. The Clark’s Cave bone deposit and the late Pleistocene paleoecology of the Central Appalachian Mountains of Virginia. Bulletin of Carnegie Museum of Natural History 2:48–51. Harvey, M.J. 1992. Bats of the Eastern United States. Arkansas Game and Fish Commission, Little Rock, AR. 46 pp. Harvey, M.J., J.S. Altenbach, and T.L. Best. 1999. Bats of the United States. Arkansas Game and Fish Commission, Little Rock, AR. 64 pp. Jones, C. 1977. Plecotus rafinesquii. Mammal Species 69:1–4. Klippel, W.E., and P.W. Parmalee. 1982. Diachronic variation in insectivores from Cheek Bend Cave and environmental change in the Midsouth. Paleobiology 8:447–458. Krebs, C.J. 2000. Ecology Methodology. 2nd Edition. Addison Wesley Logman, Inc., Menlo Park, CA. 620 pp. Kunz, T.H., and R.A. Martin. 1982. Plecotus townsendii. Mammalian Species 175:1–6. Kurta, A., and R.H. Baker. 1990. Eptesicus fuscus. Mammalian Species 356:1–10. Kurtén, B., and E. Anderson. 1980. Pleistocene Mammals of North America. Columbia University Press, New York, NY. 442 pp. Lance, G.N., and W.T. Williams. 1967. Mixed-data classificatory programs. I. Agglomerative system. Australian Computer Journal 1:15–20. Martin, P.S. 2005. Twilight of the Mammoths. University of California Press, Berkeley, CA. 269 pp. Mercer, H.C. 1894. Progress of fieldwork. Department of American and Prehistoric Archaeology of the University of Pennsylvania. Archeologist 2:117–118. Parmalee, P.W. 1961. A recent find of jaguar bone in a Tennessee cave. Journal of the Tennessee Academy of Science 37:81–85. Pielou, E.C. 1991. After the Ice Age: The Return of Life to Glaciated North America. University of Chicago Press, Chicago, IL. 366 pp. Renkonen, O. 1938. Statisch-okologische Untersuchungen uber die terrestiche kaferwelt der finnischen bruchmoore. Annales Botanici Societatis Zoologicae Botanicae Fennicae Vanamo 6:1–231. Schubert, B.W., J.I. Mead, and R.W. Graham. 2003. Ice Age Cave Faunas of North America. Indiana University Press, Bloomington, IN. 299 pp. Shump, K.A., Jr., and A.U. Shump. 1982a. Lasiurus borealis. Mammalian Species 183:1–6. Shump, K.A., Jr., and A.U. Shump. 1982b. Lasiurus cinereus. Mammalian Species 185:1–5. Simpson, E.H. 1949. Measurement of diversity. Nature 163:688. 628 Southeastern Naturalist Vol. 10, No. 4 Smith, B., and J. Wilson. 1996. A consumer’s guide to evenness indices. Oikos 76:70–82. Tuttle, M.D. 1975. Population ecology of the Gray Bat (Myotis grisescens): Factors infl uencing early growth and development. University of Kansas, Occasional Papers of the Museum of Natural History 36:1–24. Tuttle, M.D. 1976. Population ecology of the Gray Bat (Myotis grisescens): Philopatry, timing, and patterns of movement, weight loss during migration, and seasonal adaptive strategies. University of Kansas, Occasional Papers of the Museum of Natural History 54:1–38. Tuttle, M.D. 1979. Status, cause of decline, and management of endangered Gray Bats. Journal of Wildlife Management 43:1–17.