2008 SOUTHEASTERN NATURALIST 7(3):429–448 Ant Community Composition Across a Gradient of Disturbed Military Landscapes at Fort Benning, Georgia John H. Graham1,*, Anthony J. Krzysik2, David A. Kovacic3, Jeffrey J. Duda4, D. Carl Freeman5, John M. Emlen4, John C. Zak6, W. Russell Long1, Michael P. Wallace3, Catherine Chamberlin-Graham1, Jonathan P. Nutter1,7, and Hal E. Balbach8 Abstract - Military training, soil texture, and ground cover infl uence ant communities at Fort Benning, a military installation in west-central Georgia. We sampled 81,237 ground-dwelling ants (47 species in 20 genera) with pitfall traps at 40 sites on a continuum from nearly pristine forest to highly disturbed training areas. We also measured 15 environmental variables related to vegetation and soil. Sites disturbed by military training had fewer trees, less canopy cover, more bare ground, and more compact soils with shallower A-horizons than comparable undisturbed sites. Pheidole bicarinata, Dorymyrmex smithi, and Pogonomyrmex badius dominated the most highly disturbed sites. Competitively submissive myrmicines, such as Aphaenogaster and Crematogaster, and formicines, such as Camponotus and Formica, were abundant in the undisturbed sites. Solenopsis invicta occurred in all but the least disturbed sites. Ant community composition was a useful indicator of disturbance at Fort Benning. Introduction Ants are useful indicators of disturbance (Andersen 1997), and thus may help resource managers understand the effects of military training on the ecosystems that support such training. Mechanized infantry training disturbs soil and vegetation (Althoff and Thien 2005, Ayers 1994, Goran et al. 1983, Grantham et al. 2001, Haugen et al. 2003, Krzysik 1984). We know little, however, about the effects of such military training on the communities of terrestrial animals, especially invertebrates (but see Althoff and Thien 2005, Schaeffer et al. 1990, Woinarski et al. 2002). Recently, Graham et al. (2004) studied the effects of military training on ant communities of upland mixed pine-hardwood forests at Fort Benning in the Fall-line Sandhills of westcentral Georgia, a physiographic transitional area between the Piedmont and Coastal Plain. This study was part of an effort to identify indicators of ecological disturbance. Here we extend our previous study by describing 1Department of Biology, Berry College, Mount Berry, GA 30149. 2Prescott College, Prescott, AZ 86301. 3Department of Landscape Architecture, University of Illinois, Champaign, IL 61820. 4USGS Biological Resources Division, Western Fisheries Research Center, Seattle, WA 98115. 5Department of Biological Sciences, Wayne State University, Detroit, MI 48202. 6Department of Biology, Texas Tech University, Lubbock, TX 79409. 7Current address - Mercer University School of Medicine, Macon, GA 31207. 8US Army ERDC-CERL, Champaign, IL 61826. *Corresponding author - jgraham@berry.edu. 430 Southeastern Naturalist Vol.7, No. 3 ant community composition across a wider range of upland sites. The previous study examined nine sites chosen for similar physiography, vegetation, soils, and history, but having different levels of disturbance. That study was for three consecutive years, and sampled ants with pitfall traps, vegetation sweeps, and searches of tree trunks. Here we examine 40 sites (including the original nine) that vary in physiography, vegetation, soils, land-use history, and disturbance. In contrast to our previous study, these sites encompass all of the upland plant associations of Fort Benning. Employing pitfall traps, we describe ant communities across two environmental gradients—a landscape disturbance gradient and a soil texture-ground cover gradient. We treat species diversity patterns in a companion paper (J.H. Graham et al., unpubl. manuscript). We find that ant community composition in the Fall-line Sandhills is altered by landscape disturbance from military training and, to a lesser degree, is infl uenced by soil texture and ground cover. Research on ant communities of the southeastern United States has focused largely on the ants of Florida (Deyrup 1986; Deyrup et al. 1988; King 2004, 2007; King and Porter 2007; Lubertazzi and Tschinkel 2003; Van Pelt 1956, 1958). In addition, there are studies on the upland ant communities of the Coastal Plain of Mississippi (MacGown and Brown 2006), the Valley and Ridge Province of Tennessee (Dennis 1938, Martelli et al. 2004), the Piedmont of North Carolina (Carter 1962), the southern Blue Ridge Mountains (Van Pelt 1963), and the Smoky Mountains (Cole 1940). Other than our previous paper (Graham et al. 2004), none of this previous research specifically address disturbance. The Fall-line Sandhills (Keys et al. 1995) are a narrow band of rolling hills forming the physiographic ecotone between the oak-hickory-pine forest of the Piedmont (Küchler 1964, Skeen et al. 1993) and the formerly pine-dominated southern mixed hardwood forest of the Coastal Plain of the southeastern United States (Küchler 1964, Ware et al. 1993). Soils are deep, sandy, nutrient-poor, and easily eroded. The pre-settlement forest of this region was mixed pine (Pinus echinata Miller [Shortleaf Pine], P. taeda L. [Loblolly Pine], P. palustris Miller [Longleaf Pine]), and hardwoods (including Quercus spp. [oak] and Carya spp. [hickory]). Fort Benning is a 73,533-ha active military training installation that lies on the Fall-line Sandhills near Columbus, GA (32°20'N, 84°41'W). Elevations range from 61–225 m above sea level. The area has a mesothermal hot summer climate (Skeen et al. 1993). Mean summer temperature at Fort Benning is 27 °C, and mean winter temperature is 9 °C; annual rainfall is 130 cm, with 53% falling from April through October (Lozar 2004, Mason 2003). Average relative humidity for nearby Columbus, GA ranges from 59–83% in January to 57–89% in July (SERCC 2007). The soils at Fort Benning include Troup loamy sand, Ailey loamy coarse sand, Cowarts and Ailey soil, Nankin sandy loam, and Troup, Vaucluse, and Pelion loamy sand (Johnson 1983, NRCS 2007). Collins et al. (2006) and Dilustro et al. (2002) discuss upland plant communities of Fort Benning. 2008 J.H. Graham et al. 431 Upland forests at Fort Benning experience a variety of disturbances associated with military training (Garten et al. 2003). The most disturbed areas are those used for mechanized infantry training with tracked vehicles (e.g., M1A1 Abrams main battle tank, M2A2 Bradley fighting vehicle, M113 and M577 armored personnel carriers, and M88 recovery vehicle), 4-wheel-drive HMMWVs (high mobility multipurpose wheeled vehicles), and logistical support vehicles. Training in some areas has removed most of the canopy, significantly changed ground-cover composition, compacted soils, and exposed more bare ground. USAIC (2005) characterizes erosion in the mechanized training areas as “excessive.” Other areas at Fort Benning are used only by foot infantry for navigation with map and compass and have foot traffic only. Finally, parts of the installation have relatively pristine landscapes that are managed for wildlife conservation, federal- and statelisted species (e.g., Picoides borealis Vieillot [Red-cockaded Woodpecker], Gopherus polyphemus Daudin [Gopher Tortoise]), contain rare and sensitive ecosystems, or represent firing-range safety fans. Thus, a complete range of military landscape disturbance is available at Fort Benning. Methods We selected forty 4-ha sites (from a list of more than 100 potential sites), encompassing disturbed and undisturbed upland forest (Table 1; see Graham and Graham [2008] for photographs). Disturbance ranged from severe, with no canopy cover, to almost pristine, and sites had been burned from two to seven times in 23 years. Sites were deliberately chosen using eight GIS databases and extensive ground-truthing to refl ect a smooth continuum of disturbance, rather than a dichotomy of disturbed and undisturbed sites. Sites were rejected if a particular level of disturbance was already well represented. To avoid edge effects within a site, we chose plot centers and transects away from habitat boundaries. At each site, we placed four 100-m transects originating at the plot center. The first transect’s orientation (azimuth) was random. Subsequent transects were 90, 180, and 270 degrees from the first. Before field data were collected, a single experienced person (A.J. Krzysik), with 21 years experience with landscape disturbance on military installations, visually assessed disturbance to vegetation and soils at each site on an ordinal scale from 1 (most pristine) to 10 (severely disturbed). Tactical and support vehicle tracks, bivouacs, foxholes, discarded cartridge cases, MRE (meal ready to eat) wrappers, rusted debris, concertina wire, gullies, and erosion reflect military training disturbance. The disturbance classification was (as far as possible) independent of the plant community and prescribed-fire regime. Areas clear-cut within the last twenty years were avoided. Measurements of habitat variables were taken in 10 systematic random, circular quadrats (0.58 m2) along each transect (40 samples per site). Ants were sampled at the 50-meter marks of each transect (20 pitfall traps per site). 432 Southeastern Naturalist Vol.7, No. 3 In May 2003, we sampled soil A-horizon depth (mm), soil compaction (Lang penetrometer units), soil texture (% sand, silt, and clay from four 10-cm diameter x 10-cm deep cores on each transect, by the hydrometer method; Bouyoucos 1951), bare ground and litter (%), canopy cover (% with a concave spherical densiometer), forb cover (%), legume cover (%), grass cover (%), woody ground cover (%), pine seedling cover (%), fern cover (%), total ground cover (%), tree density (number per 3600 m2), and maximum basal area of trees (m2/ha with a Cruz-All basal area factor gauge). Table 1. Sample sites at Fort Benning (UTM coordinates, Zone 16, Datum NAD 83). Disturbance Site Easting Northing class Description A15 699459 3574515 2 Longleaf Pine B2 705594 3573742 1 Mixed pine-southern red oak (more deciduous) D3-1 711198 3586653 4 Mixed pine-oak-hickory D3-2 712813 3587743 5 longleaf pine D3-3 713281 3588352 4 Turkey/sand post oak-pine savanna (Loblolly and Longleaf) D6-1 712642 3588100 9 Mixed pines-oak (few trees; Loblolly dominant) D6-2 712338 3588164 3 Loblolly/Shortleaf-hardwoods D11 712256 3588385 6 Mixed pine-oak-hickory (Loblolly dominant) D15-1 712234 3588600 10 Mixed pines-oak (few trees; Loblolly dominant) D15-2 712127 3587764 7 Longleaf Pine D15-3 713420 3587398 10 Mixed pines-oak (few trees; Loblolly dominant) D15-4 713295 3586546 7 Mixed pine-oak D15-5 713320 3586352 7 Mixed pine-oak-hickory (Loblolly dominant) D15-6 713352 3586039 10 Mixed pines-oak (few trees; Loblolly dominant) D15-7 713458 3586128 8 Mixed pine-oak D16-1 714374 3586214 4 Mixed pine-oak D16-2 714236 3586098 5 Loblolly/Shortleaf-hardwoods D16-3 712983 3587307 6 Loblolly/Shortleaf-hardwoods D16-4 713482 3590705 5 White/Southern Red/Post Oak-sShortleaf/Loblolly D16-5 714024 3589871 4 Loblolly/Shortleaf-hardwoods D16-6 713806 3589721 4 Mixed pine-oak D16-7 714605 3584285 4 Mixed pine-oak-hickory (Loblolly dominant) D17 714379 3584706 6 Mixed pine-oak E5 711573 3583230 1 Oak-hickory mesic deciduous forest (diverse fl ora) F1 716040 3583605 6 Mixed pine-Southern Red Oak (more pine) F4 720091 3583300 2 Longleaf Pine H1 712522 3587761 9 Mixed pines-oak (few trees; Loblolly dominant) H2 711138 3587842 8 Mixed pines-oak (few trees; Loblolly dominant) H3 712856 3588263 9 Loblolly/shortleaf-hardwoods J4 706375 3587019 8 Mixed pines-oak (few trees; Loblolly dominant) J6 708776 3588980 4 Turkey/Sand Post Oak-pine savanna (Loblolly dominant) K13 716519 3596419 1 Turkey/Sand Post Oak-pine savanna (pure Longleaf) L1 710543 3588358 2 Loblolly/shortleaf-hardwoods L2 710451 3586075 3 Mixed pine-oak-hickory (Loblolly dominant) L3 712469 3590819 3 Mixed pine-oak M1 710361 3587062 5 Mixed pine-oak-hickory (Loblolly dominant) M2 711886 3588945 5 Mixed pine-oak-hickory (Loblolly dominant) M3 712942 3588582 5 Longleaf Pine M8 703257 3595058 3 Loblolly Pine-hardwoods (more deciduous) O10 708670 3596480 3 Loblolly Pine-hardwoods (more pines) 2008 J.H. Graham et al. 433 Because some ant species may be restricted to particular plant associations, we classified the 40 sample sites by the basal area of the 26 most abundant tree species. Trees were sampled on 4 perpendicular 100-m x 10-m strip-transects, that coincided with the 4 site transects. Trees whose centerline fell within the strip-transect were identified and measured with a 5-m fiberglass DBH tape (Forestry Suppliers Inc., No. 59571). Diameter at breast height (DBH) was recorded to 0.1 cm, and only individuals with a DBH of at least 5 cm were tallied. Fire is an important disturbance in forested ecosystems in the southeastern United States. Fort Benning Land Management Branch provided data on fire coverage by forest tract and year (1980–2003). We placed these into GIS data layers. Data for prescribed burns were available for the years 1980–1988, 1991–1994, and 1997–1999. Data for wildfires were available for the years 1980–1994 and 1997–1999. Since 2000, land managers have implemented a strategy of burning all stands in a training compartment at least once every three years. Thus, we assumed that sites scheduled for prescribed burns after 2000 were burned. We estimated fire frequency, coverage, and years since the last fire within a compartment. We were unable to determine a fire’s intensity and duration. Between 6 and 22 May 2003, we sampled ground and litter-dwelling ants with pitfall traps (9-ounce Solo plastic cups, 7 cm diameter), set out in 4 clusters (5 traps per cluster, 20 traps per site, 800 pitfalls total), at the 40 sites. Pitfalls were left out for 24 hours, because restricted access to the training sectors made it difficult to visit a site on more than two consecutive days. Maximum daily air temperature at six of our sites during the sampling period averaged 28.2 ± 3.2 °C (± SD); minimum daily air temperature averaged 17.3 ± 3.3 °C. The only significant rainfall (58.42 mm) of the sampling period occurred on the last day of sampling, after the pitfalls at two sites (O10 and M8) had been in place for nearly 24 hours. More detailed procedures for sampling and identifying ants are elaborated in Graham et al. (2004). The chief advantage of pitfalls is speed and efficiency; their chief disadvantage is that species accumulation curves for pitfalls are shallower than for some other approaches, especially with a well-developed litter community. Nevertheless, pitfalls have been used for many years to sample grounddwelling ants (Abensperg-Traun et al. 1996, Bestelmeyer and Wiens 1996, Boulton et al. 2005, Corley et al. 2006, Lassau et al. 2005, Majer and Nichols 1998, Ottonetti et al. 2006, Punttila et al. 1991, Woinarski et al. 2002). Finally, the forests at Fort Benning generally have sparse litter; because of a 3-year burn cycle, sampling the litter at most of our sites was impractical. King (2004), working in similar warm-temperate upland habitats in Florida, concluded that pitfalls were the best method when leaf litter was scarce. Thus, we have probably under-sampled many cryptic species, including the ponerines, the thief ants (Solenopsis spp.), and species of Strumigenys and Pyramica. Nevertheless, for comparing land-use impacts, pitfalls make a great deal of sense (Andersen 1991, Woinarski et al. 2002). 434 Southeastern Naturalist Vol.7, No. 3 Voucher specimens will be deposited with the Georgia Museum of Natural History in Athens, GA. The remaining specimens are deposited at Berry College. See Table 2 for a list of species. Statistical analysis Rather than relate each species of ant to each species of tree, we used multidimensional scaling (MDS; PC-ORD, Version 4, MjM Software Design, Corvallis, OR) to reduce a 26-dimensional space of tree species to a smaller dimensional space representing underlying environmental gradients. MDS amounts to an indirect ordination of species on unknown, rather than known, environmental gradients. Using the basal areas of the 26 tree species, we calculated Sorenson (Bray-Curtis) distances between all pairwise combinations of the 40 sites. These distances were then transformed into ranks. The Kruskal algorithm was then used to find a new representation of the samples in a reduced Euclidean space. A two-dimensional representation seemed best in terms of interpretability and stress (S = 14.1). To examine the relationships among soil texture, disturbance, and plant community, we estimated the Spearman ρ rank correlation between the multidimensional scaling scores, % clay, % sand, and disturbance class. Canonical correspondence analysis (CCA; MVSP, Version 3.1, Kovach Computing Services, Anglesey, Wales, UK) ordered the ant species and sites along underlying environmental gradients (Palmer 1993, Ter Braak 1986). Unlike multidimensional scaling, this is a direct ordination of species abundance on environmental gradients. To minimize the infl uence of abundant species, we log transformed the species counts (Ter Braak 1986) and downweighted rare species by their frequency. Correlations among the MDS and the CCA axes were examined by Spearman’s ρ rank correlation. Because CCA’s assumption that species distributions are unimodal may be false (see Ejrnæs 2000), we used cubic spline smoothing (PSI-Plot, Version 7.8, Poly Software International, Pearl River, NY) to further explore species distributions. Rather than use the species counts, we fit a cubic spline to species presence/absence data. Only species present in at least three of the 40 sites were included in the single-species analysis. Results We identified and counted 81,237 ants (47 species in 20 genera) in our pitfall samples (Table 2; see Graham and Graham [2008] for a list of species at each sample site). Over the entire 4-year study (including Graham et al. 2004), we have sampled 222,705 ants (52 species in 23 genera). Thus, this sample is representative of the wider, long-term study. Soils and forest community By the USDA soil classification scheme, 26 sites were loamy sands, 11 were sandy loams, and three were sands. Only 11 sites had clay content exceeding 10%, and none had clay content exceeding 13.6%. Neither sand nor clay content was correlated with disturbance (Spearman’s ρ = -0.066, n = 40, P = 0.686, and ρ = 0.210, n = 40, P = 0.193, respectively). 2008 J.H. Graham et al. 435 Table 2. List of ant species collected at Fort Benning (2003), plus species abbreviations, and canonical correspondence scores on CCA1. Negative values of CCA1 indicate the species distribution is centered in undisturbed sites; positive values indicate the species distribution is centered in disturbed sites. Subfamily Species Abbreviation CCA1 Ponerinae Hypoponera inexorata (Wheeler) Hypine -0.35 Myrmicinae Aphaenogaster fl oridana Smith Aphfl o -0.462 Aphaenogaster rudis complex Aphrud -0.687 Aphaenogaster treatae Forel Aphtre -0.56 Crematogaster ashmeadi Mayr Creash -0.462 Crematogaster lineolata (Say) Crelin -0.53 Crematogaster missuriensis Emery Cremis -0.428 Monomorium viride Brown Monvir -0.515 Myrmecina americana Emery Myrame -0.492 Myrmica latifrons Starcke Myrlat -0.188 Pheidole bicarinata Mayr Phebic 1.798 Pheidole crassicornis Emery Phecra -0.513 Pheidole davisi Wheeler Phedav -0.238 Pheidole dentata Mayr Pheden -0.523 Pheidole lamia Wheeler Phelam -0.151 Pheidole metallescens Emery Phemet -0.243 Pheidole morrisii Forel Phemor -0.666 Pheidole pilifera (Roger) Phepil -0.104 Pogonomyrmex badius (Latreille) Pogbad 1.313 Pyramica membranifera (Emery) Pyrmem -0.037 Pyramica sp. 2 Pyrsp2 0.004 Solenopsis invicta Buren Solinv 0.256 Solenopsis nr. carolinensis Forel Solcar -0.335 Temnothorax curvispinosus (Mayr) Temcur -0.071 Temnothorax texanus (Wheeler) Temtex -0.371 Temnothorax pergandei (Emery) Temper -0.307 Trachymyrmex septentrionalis (McCook) Trasep -0.146 Dolichoderinae Dorymyrmex bureni (Trager) Dorbur 0.351 Dorymyrmex grandulus (Forel) Dorgra 0.261 Dorymyrmex smithi Cole Dorsmi 1.351 Forelius pruinosus (Roger) Forpru 0.124 Formicinae Brachymyrmex depilis Emery Bradep -0.063 Brachymyrmex patagonicus Mayr Brapat 0.481 Camponotus caryae (Fitch) Camcar 0.215 Camponotus castaneus (Latreille) Camcas -0.463 Camponotus chromaiodes Bolton Camchr -0.442 Camponotus impressus (Roger) Camimp 0.004 Camponotus pennsylvanicus (De Geer) Campen -0.241 Camponotus socius Roger Camsoc -0.152 Formica pallidefulva Latreille Forpal -0.352 Formica subsericea Say Forsub -0.181 Lasius neoniger Emery Lasneo -0.179 Paratrechina arenivaga (Wheeler) Parare -0.136 Paratrechina faisonensis (Forel) Parfai -0.04 Paratrechina parvula (Mayr) Parpar -0.364 Paratrechina vividula (Nylander) Parviv 0.061 Prenolepis imparis (Say) Preimp -0.296 436 Southeastern Naturalist Vol.7, No. 3 Multidimensional scaling of 26 tree species across 40 sites revealed several trends (Fig. 1). The first axis (Axis 1) contrasts highly disturbed training areas having low basal area of all tree species (right part of the figure) with relatively undisturbed areas having high basal area of trees (left part of the figure). The second axis (Axis 2) contrasts xeric sites dominated by scrub oak-pine savannas (upper part of the figure) with mesic sites dominated by oaks and hickory (lower part of the figure). On Axis 2, the plant associations (ranked from xeric to mesic) are xeric scrub oak-pine savanna, Longleaf Pine forest, Longleaf Pine-oak forest, oak-pine forest, oak-hickory-pine forest, and mesic oak-hickory forest. Axis 1 is positively correlated with disturbance class (ρ = 0.567, n = 40, P < 0.001), but was uncorrelated with % sand (ρ = -0.090, n = 40, P = 0.579) and % clay (ρ = 0.096, n = 40, P = 0.558). Axis 2 is correlated with disturbance class (ρ = 0.337, n = 40, P = 0.033) and % sand (ρ = 0.328, n = 40, P = 0.039), but not with % clay (ρ = -0.168, n = 40, P = 0.301). Finally, all of the highly disturbed sites on Axis 1 fall in the center of Axis 2; eight of nine sites having disturbance classes of 8 or more were mixed pine-oak forest. No xeric scrub oak-pine savannas, Longleaf Pine forests, or mesic oak-hickory forests were ranked as highly disturbed. Canonical correspondence analysis Canonical correspondence analysis uncovered two environmental gradients infl uencing ant community structure. The first axis (CCA1) accounted for 23.8% of the variation in species abundances in the data. The second axis (CCA2) accounted for 7.9% of the variation. Figure 1. Multidimensional scaling of the tree community of 40 sites at Fort Benning, based upon the basal area of 26 tree species. Bubble size is proportional to disturbance class (1–10, least to most disturbed). 2008 J.H. Graham et al. 437 The first CCA axis is associated with disturbance (Fig. 2). The disturbance vector, which points strongly to the right, is nearly parallel to CCA1. Moreover, CCA1 is strongly correlated with disturbance class (ρ = 0.670, n = 40, P < 0.001). Other environmental variables associated with disturbed sites on the right of Figure 2 are vectors for % bare ground, soil compaction, and % grass cover. Environmental vectors associated with undisturbed sites on the left of Figure 2 are soil A-horizon depth, tree density, maximum basal area, and % canopy cover. Thus, large negative CCA1 is associated with undisturbed forest, whereas large positive CCA1 is associated with disturbed forest and early sere succession (former forest). The most disturbed sites were 5 sites (D15-1, J4, D15-6, D6-1, and D15-3) with compacted soils, thin A-horizons, and few trees (Fig. 3). The least disturbed sites (M8, K13, L3, and D3-3) had low compaction, deeper A-horizons, and more trees. CCA1 was also correlated with % clay (ρ = 0.430, n = 40, P = 0.006) and the first multidimensional scaling axis (ρ = 0.479, n = 40, P = 0.002), but not with % sand (ρ = -0.278, n = 40, P = 0.082) or the second multidimensional scaling axis (ρ = 0.230, n = 40, P = 0.153). The second axis was associated with soil texture. The largest vectors roughly paralleling CCA2 were those for % sand and % clay, which point in opposite directions (Fig. 2). CCA2 was positively correlated with % sand (ρ = 0.591, n = 40, P < 0.001) and negatively correlated with % clay (ρ = -0.517, n = 40, P = 0.001). Moreover, CCA2 was positively correlated with Figure 2. Habitat vectors as represented in the canonical correspondence analysis of 40 sites, 47 ant species, and 15 environmental variables at Fort Benning. 438 Southeastern Naturalist Vol.7, No. 3 the second multidimensional scaling axis (ρ = 0.632, n = 40, P < 0.001), but not with the first multidimensional scaling axis (ρ = 0.091, n = 40, P = 0.577). Vectors associated with sand content, albeit weakly, were fire coverage and frequency. Vectors associated with clay content were % pine seedlings, % forbs, % legumes, % total ground cover, and % woody cover. Sites K13, D3-3, D3-1, J6, M8, and F4 had the soils with the greatest percentage of sand, whereas E5, D16-7, H3, D15-1, D15-3, A15, D16-3, D15-6, and D16-2 had the greatest percentage of clay. E5 and D16-7 also had the greatest total ground cover, including forbs and legumes. The three ant species most associated with open, disturbed habitats were Pheidole bicarinata, Dorymyrmex smithi, and Pogonomyrmex badius (Fig. 4). Other species associated with disturbance were Brachymyrmex patagonicus, Dorymyrmex bureni, Solenopsis invicta, Dorymyrmex grandulus, Camponotus caryae, Forelius pruinosus, and Paratrechina vividula. In contrast, species associated with the least disturbed sites were Pheidole morrisii, Aphaenogaster treatae, Aphaenogaster rudis group, Crematogaster lineolata, Pheidole dentata, Myrmecina americana, and Pheidole crassicornis, all myrmicines. Ants associated with soils having higher sand content (up to 94.9% sand) were P. morrisii, Camponotus socius, D. grandulus, Aphaenogaster fl oridana, and Pheidole metallescens. Ants associated with soils having higher clay content (up to 13.6% clay) were P. bicarinata, Camponotus castaneus, Camponotus pennsylvanicus, B. patagonicus, and P. vividula. Figure 3. Sites as represented in the canonical correspondence analysis of 40 sites, 47 ant species, and 15 environmental variables at Fort Benning. See Table 1 for site descriptions. 2008 J.H. Graham et al. 439 Ants associated with sites having high total ground cover were C. castaneus, C. pennsylvanicus, A. rudis complex, C. lineolata, and Camponotus chromaiodes. In contrast, those species associated with low total ground cover were P. badius, D. smithi, D. grandulus, P. bicarinata, C. socius, and P. morrisii. Smoothing of species responses Cubic spline smoothing allowed us to examine the individual responses of 39 species, and groups of related species, to disturbance (Fig. 5). Individual responses included the following patterns (after Ejrnæs 2000): monotonic decreasing (1 species), monotonic increasing (2), unimodal symmetric (6), unimodal truncated (12), unimodal skewed (11), and bimodal (7). Only D. smithi and P. badius, which are adapted to open conditions, had monotonic increasing distributions. In contrast, only C. chromaiodes displayed a monotonic decreasing distribution. No species had a uniform distribution, though S. invicta was the closest, with uniformly high frequency at all but the least disturbed sites. Hypoponera inexorata was the only ponerine sampled in our pitfall traps. Its distribution was unimodal and relatively symmetrical about intermediate disturbance (Fig. 5A). Most myrmicine species occurred in moderately disturbed sites, and most distributions were either skewed to the right, truncated on the left, or both Figure 4. Ant species as represented in the canonical correspondence analysis of 40 sites, 47 ant species, and 15 environmental variables at Fort Benning. See Table 2 for coded abbreviations. 440 Southeastern Naturalist Vol.7, No. 3 skewed and truncated (Fig. 5A, B, C). Only three myrmicines (P. bicarinata, P. badius, and S. invicta) frequented highly disturbed sites (Fig. 5B, C). Pheidole bicarinata, however, showed a bimodal distribution; it was clearly most frequent in the most disturbed sites, but it was also found at two sites having a disturbance classification of 4. Both sites were in mixed pine-oak stands having an open canopy (32 to 40%) and erosion elements (44 to 46% bare ground). Solenopsis invicta occurred across the disturbance gradient, but was most frequent in moderately to highly disturbed habitats. All of the Dolichoderinae were more frequent in disturbed habitats, but Dorymyrmex smithi was most narrowly restricted to highly disturbed sites (Fig. 5D). In contrast, F. pruinosus and D. bureni were more widely Figure 5. Cubic spline smoothing of predicted species frequency along a disturbance gradient at Fort Benning: (A) Ponerinae (Hypoponera) and Myrmicinae (Aphaenogaster and Crematogaster), (B) Myrmicinae (Monomorium, Myrmecina, Myrmica, Pogonomyrmex, Solenopsis, Temnothorax, Trachymyrmex), (C) Myrmicinae (Pheidole), (D) Dolichoderinae, (E) Formicinae (Brachymyrmex and Camponotus), and (F) Formicinae (Formica, Paratrechina, Prenolepis). 2008 J.H. Graham et al. 441 distributed than D. smithi. Dorymyrmex grandulus was rare in our samples and was only found in moderately disturbed sites. Of the formicines (Fig. 5E, F), P. vividula was the only species more frequent in moderately to highly disturbed sites; its unimodal distribution was truncated on the right. Camponotus socius, B. patagonicus, Paratrechina parvula, and P. arenivaga were most frequent in moderately disturbed sites, but were absent from the most highly disturbed sites. Camponotus castaneus, C. chromaiodes, Formica pallidefulva, and Prenolepis imparis were more frequent in lightly disturbed sites. Discussion Ant communities of the Fall-line Sandhills at Fort Benning vary across two uncorrelated gradients, a disturbance gradient and a soil texture-ground cover gradient. Disturbance attributable to military training accounts for most of the variation among ant communities. Soil texture and ground cover account for the remaining variation. Fire, in contrast, has no discernible infl uence on ant community composition. Although we examined ant communities in a 2-1/2 week period of a single year, our study had a built-in internal control. We had already sampled nine of the 40 sites in the previous three consecutive years (Graham et al. 2004). Over the four-years (2000–2003), there was little year-to-year variation for most species despite a period of progressively increasing rainfall: extreme drought in 1999 to normal rainfall in 2003 (NOAA 2007). Only Prenolepis imparis, a species active on the surface in cool weather, and Brachymyrmex patagonicus showed much year-to-year variation in abundance. For P. imparis, the variation in the number collected was probably due to temperature variation among years. Disturbed military landscapes of Fort Benning Training with tactical vehicles does the most damage to the landscape, and it occurs primarily on the highest elevations, along fl at ridgelines. Dilustro et al. (2002) found that fl at ridgelines used for mechanized maneuvers often have the greatest sand content, while those sites used for light infantry training often have more clay (up to 48% in their study). Nevertheless, soil texture in our 40 sites was statistically unrelated to disturbance, perhaps because the range of variation was so small. Most of our sites were loamy sands. Of the sandy loams, the maximum clay content was only 13.6%. Although disturbance was unrelated to soil texture in our study, it was associated with plant community composition. The effect was most pronounced on the first multidimensional scaling axis (Axis 1). Disturbance, especially by tracked vehicles, infl uences the total basal area of all tree species. At the most disturbed sites, the forest was reduced to widely scattered clusters of trees separated by an expanse of grasses, forbs, and bare ground. The relationship of disturbance with the second multidimensional scaling axis was more a consequence of where most training occurs, on 442 Southeastern Naturalist Vol.7, No. 3 sandy ridgetops. Consequently, training was more likely to occur in mixed pine-oak forest dominated by Loblolly Pine. Soil texture (% sand) was unrelated to Axis 1 and was only moderately related to Axis 2. It was especially surprising that % clay, which is the main indicator of fertility and water holding capacity of soils (Brady and Weil 1999, Schaetzl and Anderson 2005), had so little infl uence on the plant community. This may be a consequence of limited variation; % clay ranged from only 4.2% to 13.6%. The two sites that had the greatest clay content (A15 and D16-4) had forests of Longleaf Pine and oak-pine, not mesic hardwood forest as one might expect. In fact, the mesic hardwood forest at E5 had only moderate clay content (8.7%). Percent sand, on the other hand, ranged from 57.9 to 94.9%. The two sites with the greatest sand content (87.9 and 94.9%) had scrub oak-pine savanna and oak-hickory-pine forest. Ant communities of disturbed military landscapes According to Andersen (1995, 1997), the Dolichoderinae, such as the three species of Dorymyrmex and F. pruinosus, are the behavioral dominants expected in open, warm habitats most favorable to the majority of ant species. As a group, ants are decidedly thermophilic (Andersen 1997, Brown 1973, Hölldobler and Wilson 1990), though there are many exceptions (such as P. imparis and Formica subsericea). Three species of ants were restricted to highly disturbed habitats at Fort Benning: Pheidole bicarinata, Dorymyrmex smithi, and Pogonomyrmex badius. Three others—S. invicta, D. bureni, and F. pruinosus—were also present in highly disturbed habitats, but occurred in moderately to lightly disturbed habitats as well. Solenopsis invicta, the Red Imported Fire Ant, was widely distributed across virtually the entire disturbance gradient. More submissive myrmicines (P. morrisii, A. treatae, A. rudis group, C. lineolata, P. dentata, M. americana, and P. crassicornis) lie to their left on the disturbance gradient, occupying cooler, shaded sites. Only one myrmicine (P. bicarinata) occurred primarily in open habitats, and its colonies and overall numbers were small. It was especially prevalent in the areas that had large expanses of bare soil. Our findings differ from those of Woinarski et al. (2002), who found much less modification of the ant community by military training in northern Australia. Training in their study involved armored personnel carriers, as compared to tanks and fighting vehicles in ours. By their own admission, Woinarski et al. (2002) did not sample the most disturbed sites in the military reservation, whereas we sampled a wide range of disturbance. Ant communities on a soil texture-ground cover gradient Ants often respond more to soil texture than they do to disturbance. In semiarid habitats of Colorado and New Mexico, soil texture and the associated plant community have a greater influence on ant communities than does grazing (Bestelmeyer and Wiens 2001). Moreover, Johnson (2000) found that desiccation tolerance of two species of Pogonomyrmex 2008 J.H. Graham et al. 443 was related to their distribution on a soil-texture gradient; the species on the sandier soils had greater desiccation tolerance. At Fort Benning, soil texture assumes more importance in lightly to moderately disturbed sites. Species associated with the greatest sand content are P. morrisii, C. socius, D. grandulus, A. floridana, and P. metallescens. Those species associated with the greatest clay content are P. bicarinata, C. castaneus, C. pennsylvanicus, and P. vividula. Forest composition in the Sandhills also varies with soil texture (Dilustro et al. 2002; Gilliam et al. 1993; Wells 1928, 1942). In general, hardwoods develop on soils having higher clay content, though there was much unexplained variation at our sites. Of our 40 sites, the one having the highest clay content (A15) was primarily Longleaf Pine, with a groundcover of bracken fern, suggesting a history of fire. The site having the highest sand content (J6) was a scrub oak-pine savanna. The history of a particular site (fire, last clear-cut, etc.) undoubtedly has much to do with its forest cover. The forest composition in turn infl uences parts of the ant community by infl uencing sunlight penetration and soil temperatures (J.H. Graham et al., unpubl. manuscript). Ants as indicators of military disturbance Ant community composition is a useful biological indicator of disturbance at Fort Benning. Our previous paper (Graham et al. 2004) showed that species richness of ants might be a useful indicator as well, but recent data (Graham et al., in press) suggests that species richness is greatest with intermediate disturbance. Thus, low species richness is ambiguous; it might indicate either high or low disturbance. In contrast, the presence of D. smithi, P. bicarinata, and P. badius is strongly linked to disturbance, particularly in mixed pine-oak forest. Dorymyrmex smithi, in particular, is extremely abundant in highly disturbed habitats, replacing D. bureni as disturbance increases. Conversely, the absence of many myrmicines and formicines, most notably C. castaneus, F. pallidefulva, and P. morrisii, also characterizes disturbed sites. Solenopsis invicta is a species often associated with disturbance (King and Tschinkel 2006), but at Fort Benning it was so ubiquitous as to be uninformative as an indicator species. Because mechanized training at Fort Benning is mostly confined to mixed pine-oak forest on the ridge tops, we do not know whether D. smithi, P. bicarinata, and P. badius are also the dominant ants in disturbed xeric scrub oak-pine savanna, Longleaf Pine forest, and mesic oak-hickory forest. We found only one worker of D. smithi in the three xeric oak-pine savanna sites and none in the Longleaf Pine or mesic sites. Dorymyrmex bureni, however, is present in all but the mesic sites, suggesting a habitat favorable to Dorymyrmex. Moreover, Joe MacGown (Mississippi State University, Mississippi State, MS, pers. comm.) found D. smithi to be common in disturbed Sandhill Scrub of nearby Taylor County, GA, and restricted to disturbed areas at Ohoopee Dunes Natural Area of Emanuel County, GA. Finally, he found P. badius throughout the sandhill scrub at both sites. In contrast to D. smithi and P. badius, neither Joe MacGown nor we have found evidence of 444 Southeastern Naturalist Vol.7, No. 3 P. bicarinata in disturbed (or undisturbed) sandhill scrub. The prevalence of D. smithi, P. bicarinata, and P. badius in disturbed habitats at Fort Benning does not establish these species as disturbance specialists. More likely, they are species requiring sandy soils and an open canopy. Ant communities and fire Fire had no discernable infl uence on ant community composition at Fort Benning. This lack of association between ant species and fire may have been a consequence of the three-year burn cycle, which would be expected to remove much of the leaf litter (as well as the species associated with it). Moreover, our sampling strategy may have undersampled ants of the leaf litter. On the other hand, the three-year burn cycle had only been recently instituted when we sampled ants. Not all sites had been burned within the past three years, and several sites had not been burned in more than 10 years. Furthermore, Andrew (2000) found that cryptic species in eucalyptus forests of Australia were equally common in unburned sites with litter and burned sites without litter. Conclusion Landscape disturbance to vegetation and soils from military training alters ant community composition in the southeastern Fall-line Sandhills. To a lesser degree, soil texture and ground cover are also associated with ant community composition. The species most associated with disturbance are a small subset of the 52 species at Fort Benning. These species are P. bicarinata, D. smithi, and P. badius. Moreover, a small subset of species is associated with undisturbed sites. These species are C. castaneus, F. pallidefulva, and P. morrisii. Both sets of species, used together, are reliable indicators of disturbance. More work is needed, however, on ants of disturbed xeric scrub oak-pine, Longleaf Pine forest, and mesic oak-hickory forest of the Fall-line Sandhills. Acknowledgments This research is a component of the Development of Ecological Indicator Guilds for Land Management, project CS-1114B, sponsored by SERDP (Strategic Environmental Research and Development Program), and is part of the SERDP Ecosystem Management Project (SEMP). SERDP is a joint program of the US Department of Defense, Department of Energy, and Environmental Protection Agency. Student research assistants of J.H. Graham as well as other costs were funded, in part, by Berry College. Michelle Brown and Lorence Pascoe helped with the fieldwork. In addition, we thank Pete Swiderek, Theresa Davo, and John Brent of Fort Benning, and Hugh Westbury and Patti Kosky of SERDP-SEMP for helping us to choose sites and to coordinate our work with Fort Benning Range Control. Natural Resources and Fish and Wildlife branches at Fort Benning provided the fire data. Mark Deyrup, Stefan Cover, William Mackay, James Trager, and Joshua King offered numerous suggestions on ant taxonomy, identification, and natural history. Andrea Woodward and Joe MacGown provided useful comments on the manuscript. The use of trade, firm, or corporation names in this publication is for the information and convenience 2008 J.H. Graham et al. 445 of the reader. Such use does not constitute an official endorsement or approval by the United States Department of Interior or the United States Geologic Survey of any product or service to the exclusion of others that may be suitable. Literature Cited Abensperg-Traun, M., G.T. Smith, G.W. Arnold, and D.E. Steven. 1996. The effects of habitat fragmentation and livestock-grazing on animal communities in remnants of gimlet Eucalyptus salubris woodland in the western Australian wheat belt. I. Arthropods. Journal of Applied Ecology 33:1281–1301. Althoff, P.S., and S.J. Thien. 2005. Impact of M1A1 main battle tank disturbance on soil quality, invertebrates, and vegetation characteristics. Journal of Terramechanics 42:159–176. Andersen, A.N. 1991. Responses of ground-foraging ant communities to three experimental fire regimes in a savanna forest of tropical Australia. Biotropica 23:575–585. Andersen, A.N. 1995. A classification of Australian ant communities, based on functional groups which parallel plant life-forms in relation to stress and disturbance. Journal of Biogeography 22:15–29. Andersen, A.N. 1997. Using ants as bioindicators: Multiscale issues in ant community ecology. Conservation Ecology 1:8. Available online at http://www.consecol. org/vol1/iss1/art8. Accessed September 30, 2007. Andrew, N., L. Rodgerson, and A. York. 2000. Frequent fuel-reduction burning: The role of logs and associated leaf litter in the conservation of ant biodiversity. Austral Ecology 25:99–107. Ayers, P.D. 1994. Environmental damage from tracked vehicle operation. Journal of Terramechanics 31:173–183. Bestelmeyer, B.T., and J.A. Wiens. 1996. The effects of land use on the structure of ground-foraging ant communities in the Argentine Chaco. Ecological Applications 6:1225–1240. Bestelmeyer, B.T., and J.A. Wiens. 2001. Ant biodiversity in semiarid landscape mosaics: The consequences of grazing vs. natural heterogeneity. Ecological Applications 11:1123–1140. Boulton, A.M., K.F. Davies, and P.S. Ward. 2005. Species richness, abundance, and composition of ground-dwelling ants in northern California grasslands: Role of plants, soil, and grazing. Environmental Entomology 34:96–104. Bouyoucos, G.J. 1951. A recalibration of the hydrometer method for making mechanical analysis of soils. Agronomy Journal 43:434–438. Brady, N.C., and R.R. Weil. 1999. The Nature and Properties of Soils. 12th Edition. Prentice Hall, Upper Saddle River, NJ. 881 pp. Brown, W.L. 1973. A comparison of the Hylean and Congo-West African rainforest ant faunas. Pp. 161–185, In B.J. Meggars, E.S. Ayensu, and W.D. Duckworth (Eds.). Tropical Forest Ecosystems in Africa and South America: A Comparative Review. Smithsonian Institution Press, Washington, DC. Carter, W.G. 1962. Ants of the North Carolina Piedmont. Journal of the Elisha Mitchell Scientific Society 78:1–18. Cole, A.C. 1940. A guide to the ants of the Great Smoky Mountains National Park, Tennessee. American Midland Naturalist 24:1–88. Collins, B., R. Sharitz, K. Madden, and J. Dilustro. 2006. Comparison of sandhills and mixed pine-hardwood communities at Fort Benning, Georgia. Southeastern Naturalist 5:93–102. 446 Southeastern Naturalist Vol.7, No. 3 Corley, J., P. Sackmann, V. Rusch, J. Bettinelli, and J. Paritsis. 2006. Effects of pine silvaculture on the ant assemblages (Hymenoptera: Formicidae) of the Patagonian steppe. Forest Ecology and Management 222:162–166. Dennis, C.A. 1938. The distribution of ant species in Tennessee with reference to ecological factors. Annals of the Entomological Society of America 31:267–308. Deyrup, M.A., and J. Trager. 1986. Ants of the Archbold Biological Station, Highlands County, Florida (Hymenoptera: Formicidae). Florida Entomologist 69:206–228. Deyrup, M.A., N. Carlin, J. Trager, and G. Umphrey. 1988. A review of the ants of the Florida Keys. Florida Entomologist 71:163–176. Dilustro, J.J., B.S. Collins, L.K. Duncan, and R.R. Sharitz. 2002. Soil texture, landuse intensity, and vegetation of Fort Benning, upland forest sites. Journal of the Torrey Botanical Society 129:289–297. Ejrnæs, R. 2000. Can we trust gradients extracted by detrended correspondence analysis? Journal of Vegetation Science 11:565–572. Garten, C.T., T.L. Ashwood, and V.H. Dale. 2003. Effect of military training on indicators of soil quality at Fort Benning, Georgia. Ecological Indicators 3:171–179. Gilliam, F.S., B.M. Yurish, and L.M. Goodwin. 1993. Community composition of an old-growth Longleaf Pine forest: Relationship to soil texture. Bulletin of the Torrey Botanical Club 120:287–294. Goran, W.D., L.L Radke, and W.D. Severinghaus. 1983. An overview of the ecological effects of tracked vehicles on major US Army installations. United States Army Corps of Engineers, Construction Engineering Research Laboratory, Champaign, IL. Technical Report N-142. Graham, C.C., and J.H. Graham. 2008. Photographic guide to sample sites and ant communities at Fort Benning, Georgia. Available online at http://facultyweb. berry.edu/jgraham/Fort_Benning_Sites.html. Accessed February 8, 2007. Graham, J.H., H.H. Hughie, S. Jones, K. Wrinn, A.J. Krzysik, J.J. Duda, D.C. Freeman, J.M. Emlen, J.C. Zak, D.A. Kovacic, C. Chamberlin-Graham, and H. Balbach. 2004. Habitat disturbance and the diversity and abundance of ants (Formicidae) in the Southeastern Fall-line Sandhills. Journal of Insect Science 4:30. Available online at http://insectscience.org/4.30. Accessed September 30, 2007. Grantham, W.P., E.F. Redente, C.F. Bagley, and M.W. Paschke. 2001. Tracked vehicle impacts to vegetation structure and soil erodibility. Journal of Range Management 54:711–716. Haugen, L.B., P.D. Ayers, and A.B. Anderson. 2003. Vehicle movement patterns and vegetative impacts during military training exercises. Journal of Terramechanics 40:83–95. Hölldobler, B., and E.O. Wilson. 1990. The Ants. Belknap Press of Harvard University Press, Cambridge, MA. 732 pp. Johnson, J.H. 1983. Soil survey of Muscogee County, Georgia. US Department of Agriculture, Soil Conservation Service, Washington, DC. 130 pp. Johnson, R.A. 2000. Habitat segregation based on soil texture and body size in the seed harvester ants Pogonomyrmex rugosus and P. barbatus. Ecological Entomology 25:403–412. Keys, Jr., J.A., C.A. Carpenter, S.L. Hooks, F.G. Koeneg, W.H. McNab, W.E. Rus2008 J.H. Graham et al. 447 sell, and M.L. Smith. 1995. Ecological units of the eastern United States: First approximation. Technical Publication R8-TP 21. Map (scale 1:3,500,000). US Department of Agriculture, Forest Service, Atlanta, GA. King, J.R. 2004. Ant communities of Florida’s upland ecosystems: Ecology and sampling. Ph.D. Dissertation. University of Florida, Gainesville, FL. 132 pp. King, J.R. 2007. Patterns of co-occurrence and body size overlap among ants in Florida’s upland ecosystems. Annales Zoologici Fennici 44:189–201. King, J.R., and S.D. Porter. 2007. Body size, colony size, abundance, and ecological impact of exotic ants in Florida’s upland ecosystems. Evolutionary Ecology Research 9:757–774. King, J.R., and W.R. Tschinkel. 2006. Experimental evidence that the introduced fire ant, Solenopsis invicta, does not competitively suppress co-occurring ants in a disturbed habitat. Journal of Animal Ecology 75:1370–1378. Küchler, A.W. 1964. Potential natural vegetation of the conterminous United States. American Geographical Society, Special Publication No. 36. New York, NY. Krzysik, A.J. 1984. Habitat relationships and effects of environmental impacts on the bird and small-mammal communities of the central Mohave Desert. Pp. 358–394, In W.C. McComb (Ed.). Proceedings of the Workshop on Management of Nongame Species and Ecological Communities. Department of Forestry, College of Agriculture, University of Kentucky, Lexington, KY. Lassau, S.A., G. Cassis, P.K.J. Flemons, L. Wilkie, and D.F. Hochuli. 2005. Using high-resolution multi-spectral imagery to estimate habitat complexity in open-canopy forests: Can we predict ant community patterns? Ecography 28:495–504. Lozar, R.C. 2004. SEMP historical meteorology evaluation for the area near Fort Benning, GA: 1999–2001. ERDC/CERL Technical Note 04/1. Lubertazzi, D., and W.R. Tschinkel. 2003. Ant community change across a ground vegetation gradient in north Florida’s Longleaf Pine fl atwoods. Journal of Insect Science 3:21. Available online at http://insectscience.org/3.21. Accessed September 30, 2007. MacGown, J.A., and R.L. Brown. 2006. Survey of ants (Hymenoptera: Formicidae) of the Tombigbee National Forest in Mississippi. Journal of the Kansas Entomological Society 79:325–340. Majer, J.D., and O.G. Nichols. 1998. Long-term recolonization patterns of ants in western Australian rehabilitated bauxite mines with reference to their use as indicators of restoration success. Journal of Applied Ecology 35:161–182. Martelli, M.G., M.M. Ward, and A.M Fraser. 2004. Ant diversity sampling on the southern Cumberland Plateau: A comparison of litter sifting and pitfall trapping. Southeastern Naturalist 3:113–126. Mason, J.M. 2003. Soil survey of Russell County, Alabama. United States Department of Agriculture, Natural Resources Conservation Service, Washington, DC. 202 pp. National Atmosperic and Oceanic Administration (NOAA). 2007. National Climatic Data Center. Available online at http://www.ncdc.noaa.gov/oa/ncdc.html. Accessed September 30, 2007. Natural Resources Conservation Service (NRCS). 2007. Web Soil Survey Version 1.1. Soil Survey Division, US Department of Agriculture, Washington, DC. Available online at http://websoilsurvey.nrcs.usda.gov/app/WebSoilSurvey.aspx. Accessed September 30, 2007. Ottonetti, L., L. Tucci, and G. Santini. 2006. Recolonization patterns of ants in a 448 Southeastern Naturalist Vol.7, No. 3 rehabilitated lignite mine in central Italy: Potential for the use of Mediterranean ants as indicators of restoration processes. Restoration Ecology 14:60–66. Palmer, M.W. 1993. Putting things in even better order: The advantages of canonical correspondence analysis. Ecology 74:2215–2230. Punttila, P., Y. Haila, T. Pajunen, and H. Tukia. 1991. Colonisation of clearcut forests by ants in the southern Finnish taiga: A quantitative survey. Oikos 61:250–262. Schaeffer, D.J., T.R. Seastedt, D.J. Gibson, D.C. Hartnett, B.A.D. Hetrick, S.W. James, D.W. Kaufman, A.P. Schwab, E.E. Herricks, and E.W. Novak. 1990. Ecosystem health. V. Field bioassessments for selecting test systems to evaluate military training lands in tallgrass prairie. Environmental Management 14:81–93. Schaetzl, R.J., and S. Anderson. 2005. Soils: Genesis and Geomorphology. Cambridge University Press, Cambridge, UK. 817 pp. Skeen, J.N., P.D. Doerr, and D.H van Lear. 1993. Oak-hickory-pine forests. Pp. 1–34, In W.H. Martin, S.G. Boyce, and A.C. Echternacht (Eds.). Biodiversity of the Southeastern United States: Upland Terrestrial Communities. John Wiley and Sons, Inc., New York, NY. 373 pp. Southeastern Regional Climate Center (SERCC). 2007. Historical climate data. Available online at http://www.sercc.com/. Accessed September 30, 2007. Ter Braak, C.J.F. 1986. Canonical correspondence analysis: A new eigenvector technique for multivariate direct gradient analysis. Ecology 67:1167–1179. United States Army Infantry Center (USAIC). 2005. Integrated natural resources management plan, Fort Benning Army Installation. Unpublished document. Fort Benning, GA. 344 pp. Van Pelt, A.F., Jr. 1956. The ecology of the ants of the Welaka Reserve, Florida (Hymenoptera: Formicidae). American Midland Naturalist 56:358–387. Van Pelt, A.F., Jr. 1958. The ecology of the ants of the Welaka Reserve, Florida (Hymenoptera: Formicidae). Part II. Annotated list. American Midland Naturalist 59:1–57. Van Pelt, A. 1963. High altitude ants of the southern Blue Ridge. American Midland Naturalist 69:205–223. Ware, S., C. Frost, and P.D. Doerr. 1993. Southern mixed hardwood forest: The former Longleaf Pine forest. Pp. 447–493, In W.H. Martin, S.G. Boyce, and A.C. Echternacht (Eds.). Biodiversity of the Southeastern United States: Lowland Terrestrial Communities. John Wiley and Sons, Inc., New York, NY. 502 pp. Wells, B.W. 1928. Plant communities of the Coastal Plain of North Carolina and their successional relations. Ecology 9:230–242. Wells, B.W. 1942. Ecological problems of the southeastern United States Coastal Plain. Botanical Review 8:533–561. Woinarski, J.C.Z., A.N. Andersen, T.B. Churchill, and A.J. Ash. 2002. Response of ant and terrestrial spider assemblages to pastoral and military land use, and to landscape position, in a tropical savanna woodland in northern Australia. Austral Ecology 27:324–333.