Dietary restriction (DR) lengthens life span in wide range of vertebrate and invertebrate species. The molecular mechanism by which DR increases life span and the universality of its effects (and hence its applicability to humans) are currently debated in gerontology. This article addresses these two problems from both an experimental perspective, using the nematode C. elegans as a model system, and a theoretical viewpoint, by appealing to recent mechanistic and evolutionary models of aging. Molecular mechanisms of aging are analysed by contrasting the rate of living/oxidative stress hypothesis with the metabolic stability/longevity hypothesis, a new model of aging which postulates that the robustness of metabolic networks, rather than metabolic rate per se, is the major determinant of aging. Studies of food-restricted worms are shown to be consistent with the metabolic stability/longevity hypothesis. The universality of the effects of DR is addressed in terms of directionality theory, an evolutionary model, which is based on the analytical fact that the robustness or the stability of demographic networks determines Darwinian fitness. Directionality theory, in conjunction with the metabolic stability hypothesis, predicts that DR will have negligible effects on equilibrium species (late age of sexual maturity, small size of progeny sets and broad reproductive span) and large effects on opportunistic species (early age of maturity, large size of progeny sets, narrow reproductive span). Empirical studies using C. elegans (an opportunistic species) and computational studies on human populations (an equilibrium species) are shown to be consistent with these predictions.