In each of two experiments, different groups of pigeons were required to discriminate between one of two basic kinds of stimulus differences: stimulus quality or stimulus location. For stimulus-quality groups, a key was illuminated by one of two colors on trials ending with food delivery and by the other color on trials ending with no food. For stimulus-location groups, a key was illuminated at one of two locations on trials ending with food delivery and at the other location on trials ending with no food. The birds began to respond differentially to the stimuli (i.e., peck the keys on food trials and not peck the keys on no-food trials) earlier in acquisiton if the stimulus qualities served as the signals for trial outcomes than if the stimulus locations served as those signals. The results from both experiments are consistent with predictions from a hypothesis regarding interactions among the qualities and locations of stimuli and responses (the "quality-location hypothesis"). Furthermore, the present results support other recent demonstrations of the important role that spatial relations among stimuli can play in classical conditioning.