Effects of experimental watering but not warming on herbivory vary across a gradient of precipitation.

Climate change can affect biotic interactions, and the impacts of climate on biotic interactions may vary across climate gradients. Climate affects biotic interactions through multiple drivers, although few studies have investigated multiple climate drivers in experiments. We examined the effects of experimental watering, warming, and predator access on leaf water content and herbivory rates of woolly bear caterpillars (Arctia virginalis) on a native perennial plant, pacific silverweed (Argentina anserina ssp. pacifica), at two sites across a gradient of precipitation in coastal California. Based on theory, we predicted that watering should increase herbivory at the drier end of the gradient, predation should decrease herbivory, and watering and warming should have positive interacting effects on herbivory. Consistent with our predictions, we found that watering only increased herbivory under drier conditions. However, watering increased leaf water content at both wetter and drier sites. Warming increased herbivory irrespective of local climate and did not interact with watering. Predation did not affect herbivory rates. Given predictions that the study locales will become warmer and drier with climate change, our results suggest that the effects of future warming and drying on herbivory may counteract each other in drier regions of the range of Argentina anserina. Our findings suggest a useful role for range-limit theory and the stress-gradient hypothesis in predicting climate change effects on herbivory across stress gradients. Specifically, if climate change decreases stress, herbivory may increase, and vice versa for increasing stress. In addition, our work supports previous suggestions that multiple climate drivers are likely to have dampening effects on biotic interactions due to effects in different directions, though this is context-dependent.