Variation in herbivore-mediated indirect effects of an invasive plant on a native plant.

Theory predicts that damage by a shared herbivore to a secondary host plant species may either be higher or lower in the vicinity of a preferred host plant species. To evaluate the importance of ecological factors, such as host plant proximity and density, in determining the direction and strength of such herbivore-mediated indirect effects, we quantified oviposition by the exotic weevil Rhinocyllus conicus on the native wavyleaf thistle Cirsium undulatum in midgrass prairie on loam soils in the upper Great Plains, USA. Over three years (2001-2003), the number of eggs laid by R. conicus on C. undulatum always decreased significantly with distance (0-220 m) from a musk thistle (Carduus nutans L.) patch. Neither the level of R. conicus oviposition on C. undulatum nor the strength of the distance effect was predicted by local musk thistle patch density or by local C. undulatum density (<5 m). The results suggest that high R. conicus egg loads on C. undulatum near musk thistle resulted from the native thistle's co-occurrence with the coevolved preferred exotic host plant and not from the weevil's response to local host plant density. Mean egg loads on C. undulatum also were greater at sites with higher R. conicus densities. We conclude that both preferred-plant proximity and shared herbivore density strongly affected the herbivore-mediated indirect interaction, suggesting that such interactions are important pathways by which invasive exotic weeds can indirectly impact native plants.