Food-and water-deprived pigeons keypecked for food or water reinforcement on alternate trials. Under one condition, explicit stimuli on the key provided information about the trial outcome; under another condition, only the alternation schedule provided this information. Latency and/or response rate differences between food-and water-rewarded trials emerged during both conditions. Response topography also differed on food-and water-rewarded trials. These differences, as revealed by duration and force measurements of the keypeck and by human ratings of the pecking responses as being water-or food-related, were anticipatory in nature. These results not only extend previous work on reward alternation and rewardspecific response topographies, but also have implications for theories of animal memory. In particular, these results are amenable to memory models that assume that an animal "codes" information that later must be recalled.Reward alternation studies have provided good evidence that animals can remember the reward outcome of a trial and use this information as a discriminative stimulus. Capaldi and Stanley (1963), for example, exposed rats to alternating rewarded and nonrewarded runway trials, and observed the development of an appropriate alternation in running speedthe rats ran faster on rewarded trials than on nonrewarded trials. This effect occurred even when the trials were separated by 24 h (Capaldi & Spivey, 1964).Comparable results have been found for the barpress response. Rats barpressing during alternating rewarded and nonrewarded trials developed longer latencies to respond on nonrewarded trials (Gonzalez, Bainbridge, & Bitterman, 1966;Wall & Goodrich, 1964).Apparently rats also can remember and use information about the type of reward as a discriminative stimulus. Pschirrer (1972) exposed rats to a sequence of three runway trial outcomes: either milk, pellets, and nonreward, or pellets, milk, and nonreward. Trials were separated by a 15-min intertrial interval (ITI). The rats developed equally fast running speeds on milk-and pellet-rewarded trials, and slower speeds on nonrewarded trials. Pschirrer concluded that the slower running speeds on nonrewarded trials developed because one kind of reward served as a signal of reward on the following trial and the other kind as a signal of nonreward. This interpretation received support from his second experiment, in which type of reward (milk or pellets) served as a discriminative stimulus for a right-left discrimination. Here, trials were separated by 3 min and the reward for a correct arm choice (milk or pellets) was determined randomly. Rats successfully learned to choose between a left or right alley on the basis of the type of reward received on the preceding trial.In these and other reward alternation studies, the usual behavioral measure of discrimination has been speed of running in an alley or latency to barpress. One measure that has not been studied is response topography.It has been demonstrated recently that the topography of an animal's res...