Invertebrate herbivores are ubiquitous in most terrestrial ecosystems, and theory predicts that their impact on the ecosystem should depend on the diversity and productivity of plant communities. According to the resource concentration hypothesis, the effect of herbivores should decrease with increasing plant diversity due to a lower abundance of specialist herbivores in more diverse sites. According to the hypothesis of exploitation ecosystems, the impact of herbivores should be strongest at intermediate levels of primary productivity. However, it is still difficult to draw general conclusions about the importance of invertebrate herbivores in terrestrial ecosystems. To elucidate general patterns in the relationship among invertebrate herbivory, plant diversity, and productivity, we performed a long-term herbivore exclusion experiment at 14 grassland sites in central Germany. The sites covered a wide range of plant species diversity and aboveground productivity. Over a period of five years, we used pesticides to exclude invertebrate herbivores from experimental plots both above- and belowground in a factorial design, to test the hypotheses that (1) invertebrate herbivores increase plant community diversity, but decrease plant community biomass, and that (2) the effect of herbivory on plant community biomass decreases with increasing plant diversity and is strongest at intermediate levels of primary productivity.  

Results/Conclusions

Herbivores had a positive effect on plant diversity of our grasslands as pesticide application led to a decrease in plant species richness and evenness. Exclusion of belowground herbivores shifted plant functional composition towards grasses at the expense of erosulate forbs. In general, the effects of aboveground herbivores were relatively small as compared to belowground herbivores suggesting that the latter are more important drivers of plant community composition in this ecosystem. However, effects of invertebrate herbivores on plant community biomass were not consistent across sites. Effect size depended on plant species diversity of the grasslands in the way that herbivory significantly decreased community biomass in sites of high plant diversity and increased it in sites of low diversity. In contrast to our expectation, effects of herbivory were not dependent on grassland productivity.

The close link between the effects of invertebrate herbivory on plant communities and plant species richness of the grasslands suggests that the damage caused by generalist herbivores is more pronounced than the damage due to specialist herbivores. Our study is the first one demonstrating that spatial variation in the impact of herbivores among sites of a similar plant community type is governed by plant species diversity.