Global warming is generally predicted to increase the intensity of herbivore pressure on plants. Support for this prediction often comes from short-term studies, or studies conducted in controlled laboratory settings. We examined the effect of long-term experimental warming on an aphid-sagebrush interaction (Obtusicauda coweni and Artemisia tridentata) in natural field plots in the Rocky Mountains. In no year did we find support for the prediction that warming increased aphid abundance or population growth. In fact, warming decreased aphid density on sagebrush in one year, tended to decrease aphids in a second year, and had no effect in a third year. In enclosures that excluded predators, warming decreased aphid population growth by an amount consistent with observed field density differences. Warming increased sagebrush carbon:nitrogen (C∶N) ratio and plant size, but there was no significant correlation between these variables and aphid growth or density. In a separate snow-manipulation experiment in unwarmed plots, the timing of snowmelt did not affect aphid density. In conclusion, warming reduced or did not affect aphid density in each of three years, but this effect could not be explained by differences in plant size, bulk C∶N ratio, predation, or snowmelt date. Our results suggest that long-term studies within a natural community context may provide counterexamples to the prediction that warming will increase herbivore pressure on plants.