While small-scale, plot-level field manipulations are often used to examine how global change factors affect plant productivity and relative species abundance, it can be difficult to differentiate between direct treatment effects on plants and indirect effects. We explored the question of whether changes in herbivory in response to warming and nitrogen addition would alter plant biomass in a plot-level experiment in a grass-dominated old-field. We hypothesized that plant biomass would be biased by disproportionate herbivory in the treatment plots relative to control plots. We excluded rodents from subplots of the main plots using wire mesh cylinders, and we excluded molluscs separately using copper-coated cylinders. We measured biomass 5 times between June and November, and examined differences in biomass between exclosed and open areas among warmed, fertilized and control plots.

Results/Conclusions

In heated plots, biomass in mollusc exclosures was 2 times greater than in unexclosed areas, whereas there was no exclosure effect in ambient temperature plots. In nitrogen fertilized plots, biomass in rodent exclosures was 3 times greater than in unexclosed areas, and it was only 1.2 times greater in unfertilized plots. These interactions between treatment and exclosure effects were established early in the season and did not change significantly over time. Our results demonstrate that herbivores indeed concentrated their feeding disproportionately in the warmed and fertilized plots. We suggest that the increased herbivory in treatment plots was not an accurate simulation of future herbivore responses to global change. Rather, it was an artifact of applying our treatments at the plot level.