| Abstract: | We used live-trapping and foraging experiments to describe use of roadside vegetation by kangaroo rats ( Dipodomys ordii ) in short-grass prairie in Colorado, and to determine whether this species' perception of predation risk differed between roadside habitats, where predators were expected to be most active, and locations 120 m into adjacent grazed pastures. Giving-up densities (GUDs) were measured in paired seed trays placed beneath saltbush ( Atriplex canescens ) shrubs and in the open, 3 m from shrubs along transects in roadside and pasture locations. Trials were conducted on nights with a partial (1st-quarter), full, and new moon to assess how intensity of risk affected microhabitat use. Kangaroo rats were much more abundant in areas with saltbush cover than in grazed prairie, and were 4–6 times more numerous along roadsides in saltbush areas than in other locations. On dark nights and in pasture locations, foraging rates of kangaroo rats were similar in shrub trays and open trays. The fewest seeds were removed (i.e., GUDs were highest), however, in open trays along roads on moonlit nights, suggesting that kangaroo rats recognized potential risks associated with roadsides on bright nights. The high numbers of kangaroo rats along roads suggest that benefits associated with these habitats (ease of digging, dust bathing, higher soil seed banks) exceed the costs associated with higher risk of mortality from predators and vehicles. Our results demonstrate how foraging decisions differ depending on the spatial and temporal contexts in which behaviors are measured, and underscore the potential value of less common habitats such as road margins for increasing landscape-scale diversity and wildlife habitat in grazed grasslands. |
