This study consisted of two experiments, one in rats and one in human volunteers, that used the identical progressive ratio (PR) operant procedure. In both experiments, responding was reinforced under a progressively increasing work requirement, and different groups of subjects received reinforcers that varied in sweetness. In experiment 1, rats were subjected to chronic mild stress, a well-validated animal model of depression. Performance under the PR schedule increased in subjects reinforced with conventional precision pellets (which contain 10% sucrose) or very sweet pellets, but not in subjects reinforced with sugar-free pellets. In experiment 2, volunteers were subjected to a depressive musical mood induction. Performance under the PR schedule increased in subjects reinforced with chocolate buttons, but not in subjects reinforced with with buttons made from the relatively unpalatable chocolate substitute carob. In experiment 2, depressive mood induction also increased chocolate craving, as measured by a novel questionnaire, and there were significant correlations between chocolate craving and chocolate-reinforced PR performance. These results suggest that performance under the PR schedule provides a measure of craving rather than reward, and that craving for sweet rewards is increased by depressive mood induction in both animal and human models. Implications for the interpretation of pharmacological studies using the PR procedure are also discussed.