Effects of Black-Tailed Prairie Dogs on Reptiles and Amphibians in Kansas Shortgrass Prairie

Species diversity and abundance of reptiles and amphibians were measured on and off black-tailed prairie dog (Cynomys ludovicianus) colonies to determine the extent to which herpetological species composition in a shortgrass prairie ecosystem is affected by presence of blacktailed prairie dog colonies. Ten species of reptiles and 3 species of amphibians were captured. Total amphibian and reptile abundance did not differ between prairie dog colonies and noncolonized shortgrass prairie sites, but species composition did. Mean species richness, evenness, and diversity for reptiles and amphibians were not different between treatments; however, diversity of both treatments combined was considerably higher than diversity on shortgrass prairie without prairie dogs. The mosaic pattern of prairie dog colonies enhances landscape heterogeneity and contributes to greater reptile and amphibian diversity in the shortgrass prairie biome of western Kansas. RESUMEN-Se midieron la diversidad y la abundancia de los reptiles y anfibios dentro y fuera de colonias de perritos llaneros de cola negra (Cynomys ludovicianus) para determinar hasta que punto la presencia de las colonias influye en las especies de herpetol6gicas en un ecosistema de hierba corta. Diez especies de reptiles y tres de anfibios fueron capturadas. La abundancia total de 6stos animales fue igual dentro de las colonias como en los sitios no colonizados, pero la composici6n de las especies fue diferente. El promedio de la riqueza, la diversidad, y el emparejamiento de las especies herpetol6gicas no difer6 entre los tratamientos. Sin embargo, hubo mayor diversidad en ambos tratamientos combinados que en pradera de hierba corta sin perritos llaneros. El patr6n mosaico de las colonias de perritos llaneros aumenta la heterogeneidad del paisaje y contribuye a una diversidad mas amplia de reptiles y anfibios en el ecosistema de hierba corta en el oeste de Kansas. Prior to European settlement, prairie dogs (Cynomys) were among the most numerous and widespread herbivores on North American grasslands. However, abundance and distribution of the 5 species of prairie dogs have been reduced dramatically over the past century (Summer and Linder, 1978; Miller et al., 1990). In 1902, there were an estimated 810,000 ha of black-tailed prairie dog (C. ludovicianus) colonies in Kansas (Lantz, 1903). During the period 1990 to 1992, black-tailed prairie dogs in Kansas covered only 18,843 ha, a reduction of approximately 98% from the original distribution (Vanderhoff et al., 1994). The area represented less than 0.5% of noncultivated farmland and rangeland in Kansas (Clements, 1990). Reduction of prairie dog colonies has come about largely as a result of government-supported eradication programs, habitat destruction, and disease (Cully, 1993; Miller et al., 1994). Black-tailed prairie dogs are capable of dramatically altering vegetation communities of grasslands that they occupy (Koford, 1958; Bonham and Lerwick, 1976; Coppock et al., 1983; Agnew et al., 1986; Archer et al., 1987; Weltzin et al., 1997). In mixed-grass prairies, prairie dog colonies also influence species composition and densities of mammals, birds, arthropods, and soil nematodes (O'Meilia et al., 1982; Ingham and Detling, 1984; Agnew et al., 1986). Although many studies in mixedgrass prairies have identified large numbers of animal and plant species in prairie dog colonies, there has been little work to identify the degree to which prairie species are dependent WESTERN RALIST :171-17 J E The Southwestern Naturalist on presence of black-tailed prairie dogs (Stapp, 1998; Winter, 1999). This has led to concern that the tremendous reduction in abundance and distribution of prairie dogs over the last century has put a portion of the biotic diversity of the shortgrass prairie at risk. Grazing, fire, logging, and other ecological disturbances alter abundance and diversity of reptiles and amphibians in a wide variety of habitats (Jones, 1981; Ballinger and Jones, 1985; Mushinsky, 1985; Lunney et al., 1991; Greenberg et al., 1994; Ballinger and Watts, 1995). However, there is a paucity of information regarding herpetological communities in shortgrass prairies, and little more than anecdotal reports and faunal surveys exist regarding response of reptiles and amphibians to habitat modifications by prairie dogs (Klauber, 1982; Collins and Collins, 1991; Collins, 1993). We suggest that habitat modifications caused by black-tailed prairie dogs in shortgrass prairie will favor a different suite of reptile and amphibian species than that found on unmodified prairie. Furthermore, the habitat mosaic of colonies on the prairie should yield increased diversity over shortgrass prairie habitat without prairie dogs. The purpose of this research was to determine the extent to which herpetological species composition in a shortgrass prairie ecosystem is affected by presence of black-tailed prairie dogs. METHODS AND MATERIALS-Study Area-Study sites were established on the Cimarron National Grassland in Morton Co., Kansas (37?7'30"N, 102?00'00'W), from 28 June to 6 August 1996 and 15 April to 21 July 1997. Mean annual precipitation (1901 to 1996) recorded at the Elkhart weather station in Morton Co. is 44.75 cm (National Weather Service Cooperative Observer Network, M. Knapp, pers. comm.). Precipitation varied greatly during the 2 years of this study. For the 12 months preceding the 1996 sampling period, precipitation was 34.35 cm, 77% of the long-term mean. During the 12 months preceding the 1997 sampling, precipitation was 67.97 cm, 150% of the long-term mean. Two treatments were established: in 1996, sites were selected on 5 black-tailed prairie dog colonies and 5 areas in pastures not colonized by prairie dogs (control). During the 1997 trapping season, 2 more study sites were added to each treatment, for a total of 7 study sites within each treatment. All study sites were located north of the Cimarron River on flat upland areas dominated by shortgrass prairie plant species. Study sites from each treatment were intermixed across the landscape, and were not paired. Maximum distance between any 2 study sites was 24 km. Sites occupied by prairie dogs were selected because they were among the largest prairie dog colonies on the Cimarron National Grasslands; they varied from 30 to 73 ha. Similar-sized areas were selected from control pastures without prairie dogs. Control sites were selected where vegetation was a shortgrass community type dominated by buffalograss (Buchloe dactyloides) and blue grama (Bouteloua gracilis). Study sites from both treatments had similar topography, land-use history, and soil types, which were commonly dominated by Ulysses silt loam and Richfield silt loam types (Dickey et al., 1963). In June 1997, plague (Yersinia pestis) was confirmed on 1 of the prairie dog colonies (Cully et al., 2000). Adult and juvenile prairie dogs were present there in late April, but by the end of May 1997 we were unable to locate any prairie dogs at this study site. Despite this circumstance sampling continued at this site in 1997. This study was conducted in concert with a similar study examining influence of prairie dogs on plant community structure and composition (Winter, 1999). We used a nonparametric Wilcoxon exact test (c = 0.05) to statistically test for differences in canopy cover of live grass and live forbs, canopy cover and frequency of each plant species, and percent bare ground on and off prairie dog colonies during the periods 1 July to 18 July 1996 and 18 June to 2 July 1997, and vegetation height and density during 27 May to 18 June 1997. Only values obtained from study sites that corresponded between this study and Winter's (1999) study were used to characterize vegetation differences between the two treatments. Sampling-Reptiles and amphibians were sampled using pitfall and funnel trap arrays. Trap arrays were located in each study site by randomly choosing locations on aerial photograph/grid overlays. Criteria (e.g., distance between traps; distance from study site boundary) were not used when determining trap locations. During summer 1996, 2 trap arrays were randomly located in each of the 10 study sites for a total of 20 arrays. Traps were continuously operated at each study site from 28 June through 6 August 1996. In 1997, two more study sites were added to each treatment, and the number of trap arrays was increased to 3 arrays per study site, for a total of 42 arrays. These traps were continuously operated at each study site from 15 April through 21 July 1997. Trap arrays consisted of 3 5-m drift fences, arranged 120 degrees apart, each converging at a central 5-gallon pitfall trap. Funnel traps were located at the outer ends of the drift fences. Shade boards were used over the pitfall traps and cotton sheets were wrapped around funnel traps to shelter cap172 vol. 46, no. 2 Kretzer and Cully-Effects of prairie dogs on reptiles and amphibians TABLE 1-Total number of individuals captured in trap arrays in prairie dog colonies and non-colonized control sites, Cimarron National Grasslands, Morton Co., Kansas, during the periods 28 June to 6 August 1996 and 15 April to 21 July 1997. In 1996, each treatment included 5 sites, each with 2 arrays (20 arrays total). In 1997, each treatment included 7 sites, each containing 3 arrays (42 total arrays). One each of Cnemidophorus sexlineatus, Eumeces obsoletus, and Lampropeltis triangulum was captured in the prairie dog colonies, and Tantilla nigriceps was captured once on the control plots. Prairie dog colonies Control sites Species 1996 1997 Totals 1996 1997 Totals Ambystoma tigrinum 4 3 7 0 1 1 Spea bombifrons 6 54 60 9 41 50 Bufo woodhousii 3 0 3 0 8 8 Terrapene ornata 1 4 5 0 0 0 Holbrookia maculata 5 11 16 0 1 1 Phrynosoma cornutum 0 17 17 10 21 31 Coluber constrictor 2 8* 10* 6 76* 82* Pituophis catenifer 2 6 8 3 5 8 Crotalus viridis 1 13* 14 1 * 2* * Indicates a significant difference (P < 0.05) between the control plots and the prairie dog colonies as measured by the Wilcoxon Exact Test. tured animals from the sun but still allow circulation of air through the trap. Traps were checked and captured animals were released daily. Before release, captured lizards, toads, and salamanders were marked by toe clipping as described by Martof (1953), with the exception that thumbs were not clipped on toads and long hind toes were not clipped on lizards; snakes were marked by belly scale clipping (Brown and Parker, 1976); and turtles were marked by notching marginal plates of the carapace (Cagle, 1939). Individuals that were recaptured were counted only once. Analysis-Relative abundance (number of individuals captured per 100 trap array days), species richness, Shannon's diversity index (H') and evenness measure based on Shannon's diversity index (f) were calculated for each study site. Data from 1996 and 1997 were analyzed separately, and mean values were compared statistically between treatments using a nonparametric Wilcoxon exact test (oc = 0.05) computed by SAS v. 6.11 (1996). Diversity index equations were obtained from Brower et al. (1990). Species that were captured only once were excluded from analysis. Horn's index of community similarity (Ro) was used to determine the degree of likeness between the two treatments. Horn's index varies from 0 to 1 with low values if species are mostly different, and a high value when the species composition and relative abundances are similar (Horn, 1966; Brower et