stabilized late-seral areas of the East Hills where 

 lupines are most abundant (Lesica and Cooper 

 1999). Rocky Mountain Parnassian specializes on 

 stonecrop (Sedum) and will likely be found where 

 this plant is most abundant, probably the East 

 Hills. For butterfly species requiring specific larval 

 food plants, such as some of the above, additional 

 survey work might reveal correlations between the 

 predicted distributions and various successional 

 stages of Sandhills vegetation. 



Tiger beetles provide the best examples of 

 Sandhills animal species that are closely associated 

 with early-seral vegetation and unstable sites, like 

 the 4 rare plant species studied by Lesica and 

 Cooper (1999). Indeed, tiger beetles were one of 

 the &st animal groups in which different species 

 were found to favor specific successional stages 

 (Shelford 1907). In our surveys, two of the 

 Centermial Sandhills species, Cicmdelaformosa 

 and C. decemnotata, were most common in 

 early-seral erosional or depositional sites with 

 sandy soil and low vegetation cover These 

 habitats closely match those described for C 

 formosa at other locations (Shelford 1 907, Wallis 

 1961,Knisleyl979). 



C decemnotata, however, is not considered a 

 sand dunes and blowouts specialist like C 

 formosa, but is found throughout its range in 

 upland grassland, arid shrubland, and clay banks 

 habitats (Pearson et al. 1997). For example, the 

 species is listed as an uncommon species of the 

 Idaho National Engineering Laboratory northwest 

 of Idaho Falls (Stafford et al. 1 986), where it is 

 found on sandy loams and loess (Aridisols) in 

 sagebrush-grasslands. In western Canada it has 

 been collected on dry, gravelly clay soils (Wallis 

 1 96 1 , Hooper 1 969 ) . Its broader range of habitat 

 associations might explain why C decemnotata 

 was the prevalent tiger beetle in the East Hills, 

 where the dunes are more stabilized with in- 

 creased vegetative cover (Lesica and Cooper 

 1999). Overall, however, this species too was 

 more strongly associated with low-cover sandy 

 sites in the Sandhills. 



Competition may offer further explanation for the 

 distribution patterns we found of these respective 

 tiger beetle species in the Centennial Sandhills. 

 Because the local distribution of tiger beetles is 

 determined by preferred oviposition substrates 

 (Shelford 1907), Cicmdelaformosa and C. 

 decemnotata may be competing for similar egg- 

 laying habitat in the SandhiUs. C. formosa, being 

 the larger of the two species, may be dominant at 

 breeding sites in the West Hills and exclude C 

 decemnotata fi-om the more extensive early-seral 

 habitat that is available there. Further study might 

 clarify whether such competitive interaction 

 accounts for their respective distribution patterns 

 in the Sandhills. 



Recommendations 

 and Conclusions 



Most habitats support plant and animal species 

 that are generalists-those found across a wide 

 array of habitats, and specialists-those with 

 restricted distributions and specific habitat require- 

 ments (Cody 1974). This complicates multi- 

 species management because habitat manipula- 

 tions for the benefit of one species will likely be at 

 the expense of others. The challenge, then, is to 

 design management strategies that are effective in 

 maintaining and/or restoring the full range of plant 

 and animal species native to an area. Perhaps the 

 best management approach for the Centennial 

 Sandhills fauna is to focus on disturbance pro- 

 cesses under which these species evolved, and 

 which maintain a diverse array of cover conditions 

 and stages of plant succession (Lesica and Coo- 

 per 1999) at specific spatial and temporal scales. 

 In pursuing this approach, it should be kept in 

 mind that the Sandhills themselves are but one site, 

 albeit unique, in the larger Centennial Valley 

 system. Adaptive management should include 

 larger-scale disturbances, since small-scale 

 disturbances alone probably do not replicate 

 former patch dynamics across the entire Centen- 

 nial Valley. 



18 



