In three experiments the sensitivity of instrumental responding to revaluation of the instrumental outcome in the absence of experience with the revalued outcome was examined. Hungry rats were trained to make one response for food pellets and a different response for sucrose liquid. In Experiment 1, these responses were tested in extinction when the animals were either thirsty or hungry. A significant preference for the sucrose-trained response was observed in the test conducted under thirst but not in that conducted under hunger. In Experiments 2 and 3, the effect of experience with sucrose under thirst on the magnitude of this preference was explored. Following training of the instrumental responses in Experiment 2, half of the animals received presentations of sucrose while they were thirsty; the other half received sucrose while they were hungry. In Experiment 3, the same design was used but these sucrose presentations were made contingent on an instrumental response. Also in Experiment 3, the specificity of the sucrose-response preference to a shift to thirst was examined by testing under increased and decreased levels of hunger. The results of those experiments indicated that the sucrose-response preference is exhibited only under thirst and that exposure to the sucrose under thirst only marginally enhanced that preference. These findings suggest that instrumental responding may be modified by changes in the value of its outcome in the absence of experience with the revalued outcome.