Several hypotheses predict defoliation-induced increases in individual plant fitness. In this paper we examine three such hypotheses: the Herbivore Optimization Hypothesis (HOH); the Continuum of Responses Hypothesis (CRH); and the Growth Rate Model (GRM). All three have in common predictions based on responses of defoliated individuals with the objective of explaining community and higher level phenomena. The latter two extend theory by specifying conditions for overcompensatory responses. They differ in whether overcompensation is sensitive to conditions external (CRH) or internal (GRM) to the plant. We tested these hypotheses with field experiments in a grassland system in which two native, perennial grass species replace each other along a short topographic/resource gradient. We detected positive, neutral, and negative changes in plant mass in response to partial defoliation. Patterns of responses to the edaphic and competitive environment combinations were unique to each species and neither the CRH nor the GRM were able to consistently predict responses in these grasses. Predictions of the HOH were fully supported only by the species naturally limited to lower-resource environments: overcompensation occurred in natural environments and it occurred at herbivory levels these plants experience naturally. Thus, the overcompensatory response can be important for the maintenance of local plant population distributions. However, new mechanistic theory is needed to account for the trend common to both species: overcompensatory responses to herbivory were greater in the edaphic environment in which each species was naturally most abundant.