2011 SOUTHEASTERN NATURALIST 10(4):591–608 Carabidae (Ground Beetle) Species Composition of Southern Appalachian Spruce-Fir Forests Carmen Chavez Ortiz1 and Robert A. Browne1,* Abstract - During 2003 and 2004, carabid beetles were collected from the 9 largest spruce-fir sites located in the southern Appalachian Mountains. Thirty-seven species were identified from 1817 individuals of Carabidae caught in these habitats. When adjusted for sample size, carabid beetle diversity was highest at Grandfather Mt., NC and lowest at Mt. Rogers, VA. Results from non-metric multi-dimensional scaling (NMS) ordination analysis indicate that carabid species found in spruce-fir forest cluster independently from those found in deciduous lower-elevation forests. There were no correlations between carabid species diversity and habitat area, isolation, and elevation. Introduction In the southeastern US, spruce-fir forests form a distinct ecosystem which occurs on only a few mountains in the Appalachians Mountains of Virginia, North Carolina, and Tennessee at elevations greater than 1370 m (White et al. 1993). These high-elevation areas are characterized by colder temperatures, higher humidity, and higher rainfall (1250 to 2500 mm per year) than surrounding lower elevations. During the last glacial maxima, forests dominated by Picea rubens Sargent (Red Spruce) and Abies fraseri Pursh (Poir) (Fraser Fir), covered a large portion of the southeastern US, but these species were restricted to the highest elevations as the climate warmed, and these habitats are now considered to be Pleistocene relicts (Delcourt and Delcourt 1984). During the 19th and 20th centuries, logging and associated fires reduced Southern Appalachian spruce-fir forests by as much as 90%. More recently, the widespread infestation of Adeleges picea Ratzeburg (Balsam Woolly Adelgid) has killed more than 80% of mature Fraser Fir, with additional damage from heavy metal deposition and acid rain (White et al. 1993). Since the spruce-fir forests are restricted to mountain tops, they are embedded in a matrix of lowland, primarily hardwood forests dominated by species of Quercus (Oak) and Fagus (Beech). While the differences between forest types may not always act as an absolute barrier, the changes in microhabitat are substantial and are believed to significantly limit the movement of many organisms, e.g., Desmognathus wrighti King (Pygmy Salamander; Crespi and Browne 2003), Desmognathus organi Crespi, Browne, and Rissler (Northern Pygmy Salamander; Crespi et al. 2010), and Galaucomys sabrinus Shaw (Northern Flying Squirrel; Arobogast et al. 2005, Browne et al. 1999). For these organisms, spruce-fir forests could potentially function as islands. In the current study, we assess a number of possible factors that could influence the composition and diversity of communities inhabiting nine spruce-fir “sky-island” forests in the southern Appalachians. We test the relationship between species diversity (using 1Department of Biology and Environmental Program, Wake Forest University, Winston- Salem, NC 27106. *Corresponding author - brownera@wfu.edu. 592 Southeastern Naturalist Vol. 10, No. 4 several indices) and area, isolation, and elevation, using ground beetles (Coleoptera: Carabidae) as our model taxa. Carabid beetles have proven useful in biodiversity studies since they have high species diversity and their taxonomic identification is relatively uncomplicated (Erwin 1996, Koivula et al. 2002). Relatively high numbers of individuals can be collected per unit of effort, potentially providing high statistical power for hypothesis testing. The carabid beetles in the southern Appalachians are mostly flightless and have limited dispersal capabilities (Larochelle and Lariviere 2003). Most of the carabid species found in eastern North America are generalist invertebrate predators (Larsen et al. 2003). The variety of climatic zones and complex topography of the southern Appalachians has been associated with a highly diverse biota with a relatively high level of endemics. This general observation would also apply to the Carabidae of the southern Appalachian Mountains, which form a species-rich mountain fauna (Carlton and Bayliss 2007, Darlington, 1943, Noonan et al. 1992) with numerous endemic species, especially within the tribes Trechini and Cychrini (Barr 1979, 1985; Kane et al. 1990). In the southern Appalachians, Carabidae speciation was probably enhanced by ancestral species moving into lower elevations during the Pleistocene period and then becoming isolates in higher elevations. This same process may have also contributed to the rapid speciation of other species groups in the region (Barr 1985). While there have been numerous studies of carabid species assemblage in European boreal forests (e.g., Butterfield et al. 1995; Desender et al. 1999; de Warnaffe and Lebrun 2004; Jukes et al. 2001; Kiovula et al. 2002; Niemelä 1990; Niemelä et al. 1992, 1993, 1996; Paquin 2008), fewer studies have been conducted in North America, with the majority focusing on the north temperate forests of the eastern portion of the United States (Jennings and Tallamy 2006, Larsen et al. 2003, Lenski 1982, Liebherr and Mahar 1979). Although there are substantial collection records for Carabidae in the southern Appalachians (see Barr 1979, Carlton and Bayliss 2007, Darlington 1943, Kane et al. 1990, Noonan et al. 1992) they are primarily qualitative in nature, with wide variation in search effort and seasonality. The principal objectives of this investigation were: 1) to identify carabid species found in southern Appalachian spruce-fir forest sites; 2) to measure carabid diversity via a variety of approaches for each site and for the cumulative sample; 3) to determine if species composition varied among sites; 4) to determine if carabid species found in spruce-fir forest cluster independently from those found in deciduous lower-elevation southern Appalachian forests; 5) to test for correlations between carabid species diversity and habitat area, isolation, and elevation; and 6) to determine if individual carabid species were significantly associated with wood or stone microhabitats. Materials and Methods Collection of carabid beetles We collected beetles from nine spruce-fir sites (Table 1, Fig. 1). The areas of spruce-fir forest patches included in this study ranged from 138 ha for 2011 C. Chavez Ortiz and R.A. Browne 593 Whitetop Mt. to 18,390 ha for Great Smoky Mountains. Maximum elevations of the collection sites ranged from 1344 m at Grandfather Mt. to 2008 m on Mt. Mitchell (for more detailed descriptions of each site see While et al. [1993]). Adult carabid beetles were collected during two periods: 6/21/2003–10/27/2003 and 3/20/2004–10/30/2004. Carabid beetles were collected during the day by using the active search method, i.e., manual active searching of the ground, under rocks, and under the bark of dead trees. Although some carabid species are Table I. Description of sites and collections of carabid beetles. Latitude, Elevation No. No. No. Site Code longitude (m) Area (ha) indiv. genera spp. Grandfather Mt., VA GF 36º05.958', 1324 285.73 196 8 20 81º47.740' Richland Balsam/ RIPI 35º20.700', 1874 1740.05 186 6 15 Mt. Pisgah, NC 82º57.937' Roan Mt., NC/TN RM 36º06.461', 1859 713.61 170 6 18 82º06.594' Water Rock Knob, NC WR 35º27.555', 1798 537.42 184 9 18 83º08.382' Mt. Mitchell, NC MI 35º46.149', 2008 4337.92 181 8 13 82º15.818' Mt. Rogers, VA MR 36º38.122', 1616 639.28 130 6 11 81º30.685' Whitetop Mt./ WTEG 36º38.310', 1646 138.29 161 8 18 Elk Garden, VA 81º36.373' Balsam Mt., NC BA 35º33.335', 1618 931.76 118 7 15 83º08.382' Great Smoky Mts., NC/TN GSM 1606–1920 18,390.10 491 10 22 Total 1817 13 39 Figure 1. Map of the study area with spruce-fir zones indicated as shaded areas. 594 Southeastern Naturalist Vol. 10, No. 4 nocturnal, this approach allowed us to find beetles that were resting, inactive, or hiding, and hence could include species that are often classified as nocturnal. At each site, collections occurred during at least two different calendar months between May and September. Three locations of approximately 200 m2 each were sampled within each spruce-fir site, with a total search time of 20 h at each site. Thus, the basic unit of replication was 20 h of search time for each sample site. For ordination analysis using non-metric multi-dimensional scaling (described subsequently), Carabidae were also collected from three hardwood forest sites. Two sites (KE and BF) were located in the Great Smoky Mountains National Park, adjacent to the Kephart and Bradley Fork trails (675 m and 885 m elevation, respectively), with the third site (BRCK) located in Blue Ridge National Park at Cumberland Knob (890 m elevation). Collections from the hardwood sites utilized the same techniques and were made during the same time periods as spruce-fir sites. Collected individuals were preserved in alcohol and brought to the laboratory for measurements and taxonomic identification. All collected beetles were measured for body length (tip of mandibles to apex of elytra) and identified to the species level using morphological keys (primarily Ciegler 2000). Comparisons with specimens housed at the Smithsonian National Museum of Natural History and Louisiana State University were also used in assigning species identifications. All beetles collected are stored in the Biology Department of Wake Forest University. The majority of beetles collected were larger than 3 mm in length. Since we did not collect soil samples, there is a high probability that smaller-sized beetle taxa (<3 mm), such as Trechini and Bembidiini, were underrepresented due to their small size. Therefore, our collections of carabid beetles do not represent the entire community of the family Carabidae, but instead should be considered as assessments of the carabid species >3 mm found on the forest floor up to 2 m above the forest floor. Alternative sampling methods, such as pitfall traps, nocturnal collection, fogging and vegetation beating, would probably yield different results than those we obtained (Greenslade 1964, Günther and Assmann 2004, Gutiérrez and Menéndez 1997, Lenski 1982). The goal of this study was to obtain survey data that was comparable among sample sites. Habitat variables We examined the effect of two habitat variables. 1) Area: The areas of all spruce-fir forests included in this study were measured from GIS maps using ARC View program and based on previous unpublished data. 2) Isolation: The degree of isolation among the spruce-fir forest sky-islands was estimated using three different indices. First, we measured how far north or south a spruce-fir sky-island is located in km; this served as a measure of distance from the large continuous tracks of spruce-fir forest in the northeastern US. Second, we estimated isolation based on the distance from a specific spruce-fir forest sky-island to the nearest spruce-fir forest sky-island in km. Finally, we calculated an isolation index, I, which incorporates all other patches as potential source 2011 C. Chavez Ortiz and R.A. Browne 595 populations, but is weighted by their distances and areas. This index is estimated by the following formula, where A is area and d is distance: . I = Σ(1 / dij)Aj Diversity We used several different indices to assess carabid diversity (for detailed descriptions of indices, see Magurran 2004): species richnesss (raw number of species per site), Shannon-Wiener diversity, Shannon’s evenness, Fisher’s alpha, PIE (probability of interspecific encounter), dominance (the fraction of the collection represented by the most common species), and rarefaction score. Species accumulation curves Species accumulation curves plot the number of individuals versus number of species in a sample and thus adjust species number for total sampling effort. As more individuals are collected only the rarest species are presumed to be excluded from a collection; therefore, when all species are collected, the plotted line reaches a horizontal asymptote. Smoothed species accumulation curves were constructed using EstimateS 7.52 (Colwell, 2005). For comparative purposes, and for computation of rarefaction scores, S was estimated at n = 100 individuals for each site. Data analysis The effects of area, isolation, and elevation on species richness and the other diversity indices were examined using linear regression. We created a matrix indicating how many individuals of each species were collected per site. Using this data, we constructed two different dissimilarity matrixes based on Jacard’s index of dissimilarity (J) and Sorenson’s index of dissimilarity (S) (see Magurran 2004 for detailed description of dissimilarity indices). Using both dissimilarity matrices, we compared the composition of the carabid species assemblages at the 9 spruce-fir sites and, for comparative purposes, at three hardwood forest sites using the ordination technique non-metric multi-dimensional scaling (NMS). NMS allows for differences to be easily observed graphically. We also looked at the relationship between dissimilarity in composition and geographic distance by performing Mantel tests, which calculate the correlation between the dissimilarity in the species composition between pairs of communities and the geographic distances between those pairs (Mantel 1967). For this test, we calculated the dissimilarities in the species composition using both the Sorenson’s and Jacard’s dissimilarity indices. Results Collection results A total of 1817 individuals from the 9 spruce-fir sites were identified, representing 37 species of the Carabidae family (Table 2). The Great Smoky Mts. had the highest number of individuals: 491. The other 8 sky-island sites averaged 168 ± 31 individuals. Sample sizes for the hardwood sites BF, KE, and BRCK used for 596 Southeastern Naturalist Vol. 10, No. 4 Table 2. Matrix for Carabidae taxa present/absent at nine spruce-fir sites. See Table 1 for site abbreviation codes. Average body length (in mm) with 95% confidence limit is recorded for the species at that site. A blank entry for body length measurement indicates that the species was not collected at that site. An entry of “na” for 95% confidence limits indicates that n < 3 for that species at that site. The last column indicates if there was a difference (P < 0.05) of whether that species was associated with wood, under stones, or both wood and under stones. If the column is blank, the sample size was insufficient (n < 10) to test for habitat preference. Species BA GF GSM MI MR RIPI RM WR WTEG Habitat Agonum sp. 8.50 8.13 7.80 6.35 0.31 na Atranus pubescens Dejean 8.00 na Calosoma scrutator Fabricius 22.00 na Carabus goryi Dejean 22.00 21.00 21.00 na na na Dicaelus teter Bonelli 15.00 16.00 16.33 Both na na 2.87 Harpalus spadiceus Dejean 10.50 8.00 11.00 9.00 9.00 Stone na 0.12 na 4.30 na Harpalus pensylvanicus De Geer 17.00 16.00 15.00 16.40 Both na na na na Oodes fluvialis LeConte 9.00 na Platynus angustatus Dejean 14.22 12.53 12.70 13.70 14.00 13.14 12.60 12.67 13.50 Wood 0.84 0.39 0.25 0.51 0.33 2.48 1.72 3.79 2.48 Platynus decentis Say 13.00 12.00 14.00 Both na na na Platynus tenuicollis LeConte 12.00 na 2011 C. Chavez Ortiz and R.A. Browne 597 Table 2, continued. Species BA GF GSM MI MR RIPI RM WR WTEG Habitat Tribe Pterositchini Gastrellarius honestus Say 8.06 8.06 8.14 7.98 8.23 8.10 8.02 8.58 8.02 Wood 0.11 0.12 0.03 0.06 1.11 0.50 0.42 0.59 0.30 Pterostichus acutipes Barr 14.00 Stone 0.06 Pterostichus adoxus subspecies a Say 13.07 12.50 13.48 12.75 13.06 13.47 13.70 13.26 13.11 Wood 0.44 0.27 0.31 0.16 0.12 1.52 1.21 0.54 0.77 Pterostichus adoxus subspecies b Say 11.00 13.00 14.00 Wood 1.06 0.39 na Pterostichu coracinus Newman 14.86 14.56 14.41 14.33 14.00 15.50 15.50 14.64 14.50 Both 1.12 0.79 0.18 1.43 na na 2.48 0.88 na Pterostichus lacrymosus Newman 16.00 14.46 14.36 14.92 15.00 15.00 14.00 Both na 0.51 0.49 0.63 na na na Pterostichu moestus Say 18.00 17.00 Stone na na Pterostichus mutus Say 12.33 11.00 11.00 Both 2.87 na na Pterostichus palmi Schaeffer 12.00 11.00 12.00 12.00 11.00 Stone na 1.88 0.87 na na Pterostichus relictus Newman 15.75 15.00 Stone 0.79 na Pterostichus rostratus Newman 13.86 14.33 14.07 14.33 15.00 13.17 14.00 15.00 Both 0.51 0.20 0.50 0.25 na 0.79 na na Pterostichus diligendus Chaudoir 12.00 11.00 12.50 na na na Pterostichus stygicus Say 13.00 16.00 12.50 12.00 na 0.12 na na Pterostichus superciliosus Say 16.22 1.23 598 Southeastern Naturalist Vol. 10, No. 4 Table 2, continued. Species BA GF GSM MI MR RIPI RM WR WTEG Habitat Pterostichus tristis Dejean 12.39 11.84 12.99 12.85 12.33 12.63 11.9 13.75 12.67 Wood 0.28 0.47 0.22 0.33 0.38 1.42 0.53 1.10 1.30 Pterostichus sculptus LeConte 13.00 na Tribe Cychrini Maronetus debilis LeConte 8.50 8.50 8.00 10.00 8.00 Wood na na 0.13 na na Scaphinotus andrewsii Valentine 16.00 19.50 16.73 na na na Scaphinotus elevatus Fabricius 18.00 20.27 Wood na 1.43 Scaphinotus guyoti LeConte 30.00 na Scaphinotus tricarinatus Casey 21.12 20.71 20.33 Wood 0.82 2.24 1.28 Scaphinotus viduus Dejean 30.00 27.00 29.00 Wood na na 7.45 Scaphinotus violaceus violaceus 19.75 17.00 19.00 19.00 Wood LeConte 2.38 na 2.00 na Sphaeroderus bicarinatus LeConte 16.80 17.00 17.13 Wood 0.31 na 2.48 Sphaeroderus canadensis canadensis 11.00 12.00 11.00 10.00 Wood Chaudoir 0.42 na na na Sphaeroderus canadensis lengi 11.20 11.83 11.00 11.75 11.25 Wood 0.56 0.38 na 0.80 2.47 Sphaeroderus schaumi Chaudoir 20.00 Darlington na Sphaeroderus stenostomus lecontei 16.00 14.00 16.33 15.00 15.00 15.00 Wood Dejean 2.48 na 0.51 na na na 2011 C. Chavez Ortiz and R.A. Browne 599 comparison in NMS ordination analysis were n = 21, 33, and 133, respectively. Combining spruce-fir and hardwood sites, the total number of Carabidae collected was n = 2004. Five species were collected at all spruce-fir sites: Gastrellarius honestus, Platynus angustatus, Pterostichus adoxus, Pterostichus coracinus, and Pterostichus tristis (Table 2; see Ciegler 2000 for authorities for the remaining species listed in Table 2). These 5 species were also the most abundant, accounting for 73% of all the carabids collected. When an additional species, Pterostichus rostratus, which occurred at all but one site, is added, the total increases to 83%. The majority of species (28) accounted for less than 1% of the total carabid beetles collected from all sites. For 26 species where there was sufficient sample size to test for microhabitat preference (goodness-of-fit-tests: P < 0.05), 5 species were significantly associated with stones, 14 species were significantly associated with wood (living and dead), and 7 species were found in both habitats. Every Cychrine beetle collected was associated with wood. Nearly half of all individuals for the combined spruce-fir sites were from the genus Pterostichus, with nearly 90% of all individuals from just three genera: Pterostichus, Gastrellarius, and Platynus (Fig. 2). The composition of the five most commonly found Carabidae genera differed (χ2 test: P <0.001, df = 8) among sites (Table 3). There was no significant correlation with either site latitude or longitude and the per cent composition for any of the five most common genera. Since the French Broad River has often been cited as a significant low-altitude barrier for organisms occurring in the higher altitudes of the Southern Appalachians (e.g., Browne and Ferree 2007, Crespi et al. 2003), we tested whether the Figure 2. Composition of Carabidae genera from all spruce-fir sites combined. 600 Southeastern Naturalist Vol. 10, No. 4 proportions of each of the five most common genera of Carabidae (Table 3) were significantly different between the sites south of the French Broad River (GSM, BA, WR, and RI) and the sites north of the French Broad River (MI, RM, GF, WT, and MR). A significant difference was found for Sphaeroderus (for southern sites: mean ± s.d = 1.20 ± 0.837; for northern sites: mean ± s.d. = 6.00 ± 3.367; t = 3.121, P = 0.017), but not for the other four genera. We also tested for correlation between the proportions of each genus at each site. Only the relationship between Pterostichus and Gastrellarius was significant (r = -0.831, P = 0.0055), i.e., the higher the proportion of Pterostichus at a site, the lower the proportion of Gastrellarius and vice-versa. Body length Differences in body length (Table 2, Fig. 3) occurred among the five most common genera (completely randomized design ANOVA: F = 72.82, P < 0.0001). Tukey’s tests indicate that Gastrellarius and Sphaeroderus differ significantly in length from the remaining genera, with overlap between Platynus and Pterostichus and between Scaphinotus and Sphaeroderus. For Cychrines, a distinct size progression occurs among genera, with Maronetus dominant at 8–9 mm, Scaphinotus dominant between 9–17 mm, and Sphaeroderus dominant at >17 mm. Diversity indices As estimated by raw species counts, the Great Smoky Mts. had the highest number of species from a spruce-fir sky-island (Table 4). However, after adjusting for sample size, other diversity measures ranked Grandfather Mt. as the most diverse site. Grandfather Mt. was the most diverse site as measured by the Shannon-Weiner index and Fishers’ alpha (2.33 and 5.67, respectively) and had the highest evenness value (0.78). Consistent with high evenness, Grandfather Mt. also had the lowest dominance value, with the most abundant species accounting for 27% of the total number of individuals collected. Grandfather Mt. also presents the highest probability of encountering two different species in any random pair as determined by the PIE diversity index, with a probability Table 3. Composition by genera for Carabidae from nine southern Appalachian spruce-fir forest sites. See Table 1 for site abbreviations. PT = Pterostichus, GA = Gatrellarius, PL = Platynus, SC = Scaphinotus, SP = Sphaeroderus, HA = Harpalus, Ag = Agonum, MA = Maronetus, CA = Carabus, RE = remaining genera. Genus Site PT GA PL SC SP HA AG MA CA RE GF 42 47 3 4 3 <1 0 <1 <1 <1 RIPI 51 31 2 7 8 <1 0 0 0 0 RM 24 69 1 5 <1 <1 0 0 0 0 WR 81 2 13 2 1 <1 <1 <1 0 <1 MI 78 8 10 <1 <1 1 0 <1 2 0 MR 26 32 5 34 2 0 0 0 <1 <1 WTEG 54 35 4 <1 2 4 0 0 0 <1 BA 67 14 7 0 8 <1 2 1 0 0 GSM 38 24 21 7 7 <1 2 <1 0 <1 2011 C. Chavez Ortiz and R.A. Browne 601 value of 0.87. In contrast, Mt. Rogers was the least diverse site and also had the lowest evenness value. For all locations, the species accumulation curves (Fig. 4) do not reach an obvious inflection point, indicating that increasing sampling size would continue to add new species to the survey for each site. The site with the largest sample size (the Great Smoky Mts.) continues to add new species, even for a sample size of approximately 500 individuals. Rarefaction scores (Table 4), based on 100 individuals collected from each site, show a range of species diversity estimates among sites, with the highest value at Grandfather Mt. approximately twice as large as the lowest value at Mt. Rogers. Figure 3. Profile of body lengths for the five most common genera of Carabidae collected from all spruce-fir sites. The width of the figure is proportional to the proportion of individuals with that body length. Table 4. Diversity indices for carbid beetle assemblages of spruce-fir sites. See Table 1 for site abbreviations. See text for definition of terms. Dom. = dominance. No. No. Shannon- Fisher’s Probability Rare- Site individ. spp. Weiner Evenness Alpha of encounter Dom. faction GF 185 20 2.329 0.778 5.69 0.870 0.265 16.220 MR 129 11 1.057 0.441 2.87 0.470 0.713 9.532 BA 118 15 2.066 0.763 4.55 0.814 0.373 14.448 WTEG 160 17 1.850 0.640 5.20 0.755 0.350 15.343 MI 204 12 1.760 0.686 3.09 0.778 0.358 10.304 WR 183 17 2.041 0.706 4.94 0.807 0.301 15.066 RM 169 18 2.247 0.778 5.09 0.849 0.325 15.038 RIPI 194 14 1.733 0.640 3.79 0.721 0.479 12.266 GSM 501 21 2.266 0.713 5.25 0.849 0.271 15.478 602 Southeastern Naturalist Vol. 10, No. 4 Factors influencing diversity In order to determine patterns in species composition of carabid beetles in the spruce-fir forest sites, we correlated each of the diversity indices with area and elevation. For area, there was a positive but non-significant relationship between absolute number of species and spruce-fir area (r2 = 0.345, P = 0.09). When adjusted for samples size via rarefaction analysis, the correlation between species and area was lower (r2 = 0.167, P = 0.73). There were no significant correlation between any of the diversity measures and elevation of each spruce-fir site (e.g., the correlation between species number and elevation was r2 = 0.073, P = 0.48). Carabid species composition We investigated carabid beetle species composition via ordination and by Mantel’s test. Ordination. In plots of the NMS analysis based on both the Jacard’s and Sorenson’s dissimilarity matrices (Fig. 5), the spruce-fir forest sites form a distinct cluster, indicating similarity in composition. The composition of Carabid beetles in the three hardwood forest sites (BF, KE, and BRCK), are distinct from sprucefi r forest communities. Mantel’s test. Even if isolation does not significantly affect species richness, it may influence species composition. If the beetles disperse among patches, then nearby patches would be expected to be more similar than patches that are more distant. Mantel’s tests, based on both the Jacard’s and Sorenson’s dissimilarity Figure 4. Species accumulation curves used to calculate rarefied species richness. See Table 1 for site abbreviations. 2011 C. Chavez Ortiz and R.A. Browne 603 indexes, found no significant relationship between geographic distance and community composition (based on Sorenson’s index: r = -0.00008, P = 0.48; based on Jacard’s index: r = -0.003, P = 0.46). Discussion We collected beetles using the active search method. The most commonly used alternative method, pitfall traps, has been demonstrated to have several limitations. For example, pitfall traps are limited in the area sampled. Pitfall traps are also criticized for giving non-representative samples, since they are potentially biased by differences in activity among beetle species and may also be sensitive to environmental variation such as the structure of the forest floor (Adis 1979). In contrast, active searching has the advantage of being selective for the target organisms and also allows collection of beetles inhabiting other microhabitats that are not on the ground (such as the bark of dead trees, the base of living trees with loose bark, and cracks in rocks). However, active searching also has its limitations. For example, the historical records indicate that the complete carabid beetle community also includes a large number of smaller species, such as those of the genus Trechus (Noonan et al. 1992), which are primarily found in soil or soil litter. In future studies investigating carabid beetle communities of the spruce-fir forests, a more complete record would be obtained by using a combination of sampling techniques (Larochelle and Lariviere 2003). The collections analyzed in this study included 37 species of carabid beetles. Based on the records of Darlington (1943) and Barr (1979, 1985), the most common species from our collections were also present in their collection. Since abundance data were not provided in their publications, we were Figure 5. Non-metric multidimensional scaling (NMS) of Jacard’s dissimilarity index based on species presence/absence. Sites enclosed within the ellipse are southern Appalachian spruce-fir forest sites; sites outside the ellipse are southern Appalachian hardwood forest sites. 604 Southeastern Naturalist Vol. 10, No. 4 unable to investigate if the pronounced environmental changes that have occurred in the southern Appalachians over the past several decades have led to corresponding changes in carabid beetle diversity and composition. More recent information provided by the All Taxa Biological Inventory (ATBI) of the Great Smoky Mountains National Park reports the presence of all the beetles collected in our study (Carlton and Bayliss 2007). The number of endemic species found in this study is substantially reduced in comparison with the ATBI list, but it is important to note that the ATBI includes smaller-sized species and also encompasses a vast area of hardwood forest and lower-elevation forest; sprucefir forest accounted for only a minimal part (<5%) of the whole area sampled in the GSM. With the ATBI data, there is again insufficient information on abundances to allow for quantitative comparisons. Since the spruce-fir habitats are often separated by large distances and surrounded by distinctive types of habitat, they can be considered functionally isolated. Accordingly, we hypothesized that carabid beetle diversity might be affected by the same processes that affect islands, as described in Macarthur and Wilson’s (1967) Theory of Island Biogeography. Namely, we predicted that diversity would increase with area and decrease with isolation. The Theory of Island Biogeography has previously been applied to other non-island systems including forest fragments, caves, and mountaintops, which could function as “sky-islands” (Brown 1971, Browne and Ferree 2007). Contrary to expectations, we did not find any relationships between diversity and either area or isolation. The Great Smoky Mts. is the largest continues patch of spruce-fir forest and did have the largest number of species, but this result can be attributed to the greater collection effort and greater number of individuals collected at this site. When the number of individuals was controlled for, via the use of other diversity measures such as Fisher’s Alpha, PIE, and rarefaction score, no significant relationship was found between diversity and area. There was also no relationship between any of the diversity indexes and any of the measures of isolation. Likewise, there was no relationship between dissimilarity in species composition and distance between patches (as analyzed with Mantel’s test). The lack of an isolation or distance effect indicates that either carabid beetles are dispersing equally between all sites regardless of distance or that carabid beetles are not significantly dispersing and similarity in species composition is primarily due to similar extinction trajectories after isolation. Given the degree of isolation between many of the sites, species composition based primarily on differential extinctions and not due to colonization or dispersal seems more likely. These results are consistent with other studies examining mountaintop communities. Brown (1971) tested MacArthur and Wilson’s (1967) theory for the small-mammal communities (excluding bats) inhabiting the isolated peaks of the Rocky Mountains in Nevada. Brown found a steeper species-area curve than typical of insular biotas and did not find any relationship between the numbers of species and isolation. He concluded that colonization occurred during the Pleistocene and that the current communities are relicts. He also concluded that the extinction rate has been low and the immigration rate approaches zero (Brown 2011 C. Chavez Ortiz and R.A. Browne 605 1971). In consequence, the study system is not at equilibrium (i.e., colonization does not equal extinction) as predicted by MacArthur and Wilson. The data from this study indicate that body length for the five most common genera in the spruce-fir habitats ranged from 8 mm for Gatrellarius up to 34 mm for Sphaeroderus. These 5 dominant carabid genera concur in their strong preference for forests with shaded understories and moderately moist soil covered with thick leaf litter. These species are also nocturnal and generally take shelter during the day under the loose bark of trees (either fallen or standing) or under logs and stones. They are almost all flightless, are moderate runners, and frequently climb trees. When disturbed, adults of these species emit chemicals from the pygidial glands as a defense mechanism. Although individuals of all five genera share many traits, including that they are all invertebrate predators, there is specialization within that broad category. Scaphinotus and Sphaeroderus are the dominant genera of tribe Cychrini, a highly distinct guild due to their particular body shape and their specialized feeding habitats, preying almost exclusive on snails and slugs. Their mouth parts are highly adapted to their prey, showing an elongated mandible and “spoon-like” palpomeres (Arnett and Thomas 2001, Larochelle and Lariviere 2003). Their body architecture is also structured in such a way that the pronotum and head are flexible enough to enable the beetle to get inside of the snail shell and reach the prey easily. The body-length profiles indicate that there are distinct differences in length between the larger Sphaeroderus and the relatively smaller Scaphinotus. The remaining genus within the Cychrini, Maronetus, has a body length that ranges from 7.7 to 10.2 mm. Thus, the three genera of Cychrines have a nearly step-wise progression in length from Maronetus to Scaphinotus to Sphaeroderus. For the three most common non-cychrine genera, Gastellarius has minimal overlap with any of the four genera. Platynus and Pterostichus do have some size overlap, although the largest number of Platyni are approximately 12 mm versus approximately 14 mm for Pterostichus. Although Platynus and Pterostichus are described as generalized invertebrate predators (Ciegler 2000), they may specialize in different prey items, may have different temporal preferences (seasonal or circadian), or may occur in different habitats. This study has limitations and should be considered as an initial attempt at examining the carabid species assemblages of the southern Appalachian spruce-fir forests. For all locations, the species accumulation curves do not reach an obvious inflection point, indicating that increasing sampling size would continue to add new species to the survey for all sites. Although sampling occurred during at least two different calendar months between May and September, additional sampling (e.g., monthly over at least one year) would be needed to detect seasonal specialists. Additional years of collection would be required to increase the probability of detecting rare species and to estimate annual turnover rates in the species composition at each site and for the meta-population. Our hand-collection technique almost certainly had biases (e.g., towards larger size, more active individuals, and individuals found from 0 to 2 m from the ground). Additional collection techniques are thus needed for comprehensive sampling (Liu et al 2007). 606 Southeastern Naturalist Vol. 10, No. 4 Acknowledgments We thank the Great Smoky Mountains National Park, Blue Ridge National Park, Pisgah National Forest, Jefferson National Forest, and Mount Mitchell State Park for collection permits. We also thank Terry Erwin, Chris Carlton, and Victoria Bayless for taxonomic assistance and Ken Feely for assistance with data analysis. Financial assistance was provided by the Wake Forest Environmental Program and the Science Research Fund. Literature Cited Adis, J. 1979. Problems of interpreting arthropod sampling with pitfall traps. Zoologischer Anzeiger 202:177–184. Arbogast, B.S., R.A. Browne, P.A. Weigl, and J.G. Kanagy. 2005. Conservation genetics of endangered flying squirrels of the Appalachian mountains of eastern North America. Animal Conservation 8:123–133. Arnett, R.H., and M.C. Thomas. 2001. American Beetles. CRC Press, Boca Raton, FL. Barr, T.C.J. 1979. Revision of Appalachian Trechus (Coleoptera: Carabidae). Brimleyana 2:29–75. Barr, T.C.J. 1985. Pattern and process in speciation of Trechine beetles in eastern North America (Coleoptera: Carabidae: Trechinae). Pp. 371–372, In G.E. Ball (Ed.) Taxonomy, Phylogeny, and Zoography of Beetles and Ants. Dr. W. Junk Publishers, Amsterdam, Netherlands. Brown, J.H. 1971. Mammals on mountaintops: Non-equilibrium insular biogeography. American Naturalist 105:467–478. Browne, R.A., and P.M. Ferree. 2007. Genetic structure of southern Appalachian spruce-fir “sky-island” populations of the Red-backed Vole. Journal of Mammalogy 86:104–122. Browne, R.A., S.M. Steele, P.D. Weigl, J. Kelly, and E. Eagleson. 1999. Mountain tops as islands: Genetic variation in Northern Flying Squirrel populations. Pp. 205–214, In R. Eckerlin (Ed.). Proceedings of the Appalachian Biogeography Symposium. Virginia Museum of Natural History Publications, Martinsville, VA. Butterfield, J., M.L. Luff, M. Baines, and M.D. Eyre. 1995. Carabid beetle communities as indicators of conservation potential in upland forests. Forest Ecology and Management 79:63–77. Carlton, C., and V. Bayless. 2007. Documenting beetle diversity in Great Smoky Mountains National Park: Beyond the halfway point. Southeastern Naturalist 6:183–192. Ciegler, J.C. 2000. Ground beetles and wrinkled bark beetles of South Carolina (Coleoptera: Geadephaga: Carabidae and Rhysodidae). South Carolina Agriculture and Forestry Research System, Clemson University, Clemson, SC. Colwell R.K. 2005. EstimateS: Statistical estimation of species richness and shared species from samples. Version 7.5. User’s Guide and application. Available online at http://purl.oclc.org/estimates. Accessed 3 July 2010. Crespi, E.J., L.J. Rissler, and R.A. Browne. 2003. Testing Pleistocene refugia theory: phylogeographical analysis of Desmognatus wrighti, a high-elevation salamander in the southern Appalachians. Molecular Ecology 12:969–984. Crespi, E.J., R.A. Browne, and L.J Rissler. 2010. Taxonomic revision of Desmognathus wrighti. Herpetologia 66:283–295. Darlington, P.J. 1943. Carabidae of mountains and islands: Data on the evolution of isolated faunas and on atrophy of wings. Ecological Monographs 13:37–61. 2011 C. Chavez Ortiz and R.A. Browne 607 Delcourt, H., and P. Delcourt. 1984. Late Quaternary history of the spruce-fir ecosystem in the southern Appalachian Mountains. Pp. 22–35, In P. White (Ed.). The Southern Appalachian Spruce Fir Ecosystem: Its Biology and Threats. National Park Service, Southeast Region, Gatlinburg, TN. Desender, K., A. Ervynck, and G. Tack. 1999. Beetle diversity and historical ecology of woodlands in Flanders. Belgian Journal of Zoology 129:139–156. de Warnaffe, G.B. and P. Lebrun. 2004. Effects of forest management on carabid beetles in Belgium: Implications for biodiversity conservation. Biological Conservation 118:219–234. Erwin, T.L.1996. Biodiversity at its utmost: Tropical forest beetles. Pp. 27–40, In M.L. Reaka-Kudla, D.E. Wilson, and E.O. Wilson (Eds.). Biodiversity II: Understanding and Protecting our Biological Resources. Joseph Henry Press, Washington, DC Greenslade, P.J.M. 1964. Pitfall trapping as a method for studying populations of Carabidae (Coleoptera). Journal of Animal Ecology 33:301–310. Günther, J., and T. Assmann. 2004. Fluctuations of carabid populations inhabiting ancient woodland (Coleoptera: Carabidae). Pedobiologia 48:159–164. Gutiérrez, D., and R Menéndez. 1997. Patterns in the distribution, abundance, and body size of carabid beetles (Coleoptera: Caraboidea) in relation to dispersal ability. Journal of Biogeography 24:903–914. Jennings, V.H., and D.W. Tallamy. 2006. Composition and abundance of grounddwelling Coleoptera in a fragmented and continuous forest. Environmental Ecology 35:1550–1560. Jukes, M.R., A.J. Peace, and R. Ferris. 2001. Carabid beetle communities association with coniferous plantations in Britain: The influence of site, ground vegetation, and stand structure. Forest Ecology and Management 148:271–286. Kane, T.C., T.C. Barr, and G.E. Stratton. 1990. Genetic patterns and population structure in the Appalachian Trechus of the vandykei group. Brimleyana 16:133–150. Koivula, M., J. Kukkonen, and J. Niemelä. 2002. Boreal carabid-beetle (Coleoptera, Carabidae) assemblages along the clear-cut-originated succession gradient. Biodiversity and Conservation 11:1269–1288. Larochelle, A., and M.C. Lariviere. 2003. A Natural History of the Ground-Beetles (Colopetera: Carabidae) of America North of Mexico. Pensoft, Sofia, Bulgaria. Larsen, K.J., T.T. Work, and F.F. Purrington. 2003. Habitat-use patterns by ground beetles (Coleoptera: Carabidae) of northeastern Iowa. Pedobologia 11:288–299. Lenski R.E. 1982. The impact of forest cutting on the diversity of ground beetles (Coleoptera: Carabidae) in the southern Appalachians. Ecological Entomology 7:385–390. Liebherr, J., and J. Mahar. 1979. The carabid fauna of the upland oak forest in Michigan: Survey and analysis. Coleopterists Bulletin 33:183–197. Liu, Y.J., C.L. Axmacher, C. Li, and Z. Wang. 2007. Ground beetle (Coleoptera: Carabidae) inventories: A comparison of light and pitfall trapping. Bulletin of Entomological Research 97:577–583. MacArthur, R.H., and E.O. Wilson. 1967. The Theory of Island Biogeography. Princeton University Press, Princeton, NJ. Magurran, A.E. 2004. Measuring Biological Diversity. Blackwell Publishing, Malden, MA. Mantel, N. 1967. The detection of disease clustering and a generalized regression approach. Cancer Research 27:209–220. Niemalä, J. 1990. Spatial distribution of carabid beetles in the southern Finnish taiga: The question of scale. Pp. 143–155, In N.E. Stork (Ed.). The Role of Ground Beetles in Ecological and Environmental Studies. Intercept Limited, Andover, Hampshire, UK. 608 Southeastern Naturalist Vol. 10, No. 4 Niemelä, J., Y. Haila, E. Halme, T. Pajunen, and P. Punttila. 1992. Small-scale heterogeneity in the spatial distribution of carabid beetles in the southern Finnish taiga. Journal of Biogeography 19:173–181. Niemelä, J., D. Langor, and J.R. Spence. 1993. Effects of clear-cut harvesting on boreal ground-beetle (Coleoptera: Carabidae) in western Canada. Conservation Biology 7:551–561. Niemelä, J.K., Y. Haila, and P. Punttila. 1996. The importance of small-scale heterogeneity in Boreal forests: Variation in diversity in forest-floor invertebrates across the succession gradient. Ecography 19:352–368. Noonan, G.R., G.E. Ball, and N.E. Stork. 1992. The Biogeography of Ground Beetles of Mountains and Islands. Intercept Press, Andover, Hampshire, UK. Paquin, P. 2008. Carabid beetle (Coleoptera: Carabidae) diversity in the Black Spruce succession of eastern Canada. Biological Conservation 14:261–275. White, P.S., E. Buckner, J.D. Pitillo, and C.V. Cogbill. 1993. High-elevation forests: Spruce-fir forest, northern hardwood forest, and associated communities. Pp. 305– 338, In W.H. Martin (Ed.). Biodiversity of the Southeastern United States: Upland Terrestrial Communities. John Wiley and Sons, New York, NY.