Multispecies interactions and lifetime fitness of the invasive plant Centaurea solstitialis

Background/Question/Methods Plants interact with a multitude of species, some of which enhance fitness (e.g., pollinators) while others reduce it (e.g., herbivores and pathogens). Nonadditive effects can arise when interacting with one species modifies a trait in the plant (e.g., phenology, morphology, biochemistry) that mediates a subsequent interaction. Nonadditive effects are likely to be common in nature, although they remain poorly understood. Here we examine how sequential interactions with two biological control enemies affect lifetime fitness of the invasive plant Centaurea solstitialis in a field experiment. The foliar pathogen Puccinia jaceae var. solstitialis was approved for release in 2005 and it is unclear how it will interact with the established insect biocontrol agent Eustenopus villosus, a weevil that attacks inflorescences. In 2006 and 2007 we experimentally infected naturally recruiting plants in an established C. solstitialis invasion where Eustenopus is common.

Results/Conclusions

Puccinia reduced plant height, biomass, and number of buds and inflorescences. Eustenopus showed no preference for either infected or uninfected plants, but because infection reduced the total number of buds and inflorescences, the weevil fed on a greater proportion of buds and oviposited eggs into a greater proportion of inflorescences on Puccinia-infected plants as compared to control plants. This suggests that in combination, these biological control agents have a greater than expected negative effect on the plant, which is the desired outcome in a biocontrol-weed system. However, when we estimated the proportion of seeds consumed by the Eustenopus larvae, we found that in 2006 (typical rainfall) seed production was reduced by 22% (±4% SE) in control plants but was reduced by only 3% (±10% SE) in Puccinia-infected plants. In 2007 (pronounced drought) all plants suffered higher proportional losses to seed-feeding larvae than in the previous year, but control plants were more adversely affected than infected plants (63% ±9% SE and 39% ±14% SE, respectively). Drought also reduced the direct effect of the pathogen on the plant, which in turn reduced the magnitude of the indirect interaction between the pathogen and the weevil.

The enemy enhancement approach to biological control assumes that multiple enemies attacking a plant will have additive or synergistic effects. Our data suggest that these interactions may be more complex than previously appreciated and can vary significantly from year to year and possibly site to site in response to differences in climate.