Herbivores are credited with driving the evolutionary diversification of plant defensive strategies over macroevolutionary time. For this to be true, herbivores must also cause short-term evolution within plant populations, but few studies have experimentally tested this prediction. We addressed this gap using a long-term manipulative field experiment where exclosures protected 22 plant populations from natural rabbit herbivory for <1 to 26 years. We collected seeds of Rumex acetosa L. (Polygonaceae) from our plots and grew them in a common greenhouse environment to quantify evolved differences among populations in individual plant growth rate, tolerance to herbivory, competitive ability, and the concentration of secondary metabolites (tannins and oxalate) implicated in defense against herbivores. In 26 years without rabbit herbivory, plant growth rate decreased linearly by 30%. We argue that plant growth rate has evolved as a defense against intense rabbit herbivory. In contrast, we found no change in tolerance to herbivory or concentrations of secondary metabolites. We also found no change in competitive ability, suggesting that contemporary evolution may not feed back to alter ecological interactions within this plant community. Our results combined with those of other studies show that the evolution of gross morphological traits such as growth rate in response to herbivory may be common, which calls into question assumptions about some of the most popular theories of plant defense.