COS 180-4 - Resource availability mediates the strength of density-dependent herbivory

Background/Question/Methods

Central to ecology is understanding how strongly herbivores impact plant fitness. Herbivory is highly variable at local and large spatial scales, and most studies do not direct assess fitness consequences. At local scales (e.g., patch or population level), density can be an important predictor of herbivore pressure leading to a negative (sensu resource concentration hypothesis) or positive (sensu resource dilution hypothesis) density dependent effect on herbivore damage rates. At large spatial scales, climate or resource availability may be important predictors of herbivore pressure. For instance, herbivore pressure is expected to increase with resource availability (sensu resource availability hypothesis). Although many studies have shown patterns of herbivore pressure to seedlings, saplings, and post-dispersal seed predation at different scales, few have considered the effects of conspecific density on pre-dispersal seed predation along a gradient of resource availability. Using Monarda fistulosa as a focal plant species, we sampled plots in eight populations in each of two different resource regions: a low-resource region in Montana with cooler temperatures and low summer precipitation, and a high-resource region in Wisconsin with warmer temperatures and 3-fold greater precipitation. In plots within each population, we estimated density of flowerheads, seed head damage, seed loss, and seed set.

Results/Conclusions

Our results showed that resource availability mediates the strength of density dependence with different patterns of herbivore pressure for the probability of damage and seed loss. At local scales, flowerheads density was positively correlated with the probability of damage per plot in the low resource region, which is in accordance with the resource concentration hypothesis. However, in the high resource region, we did not find any influence of flowerheads density on the probability of damage. We observed that denser plots led to a dilution in the percentage of seed loss in the high resource region, which is consistent to the resource dilution hypothesis. On the other hand, denser plots did not influence the seed loss in the low resource region. Finally, we found a greater number of seeds produced in the high resource region than in the low resource region. Our study suggests a predictable influence of conspecific density (local scale) and environmental context (large scale) on pre-dispersal seed predation with greater negative impacts in low-resource regions. These results improve understanding of how local context-dependent factors vary between resource regions highlighting different effects on the probability of damage and ultimately on fitness consequences (i.e., seed loss).