Lever presses by two rhesus monkeys produced food pellets that were assigned by both an ascending and descending series of fixed-interval schedules whose values varied between 1 and 512 sec. The amount of schedule-induced drinking was bitonically related to interreinforcement interval, reaching a maximum at approximately 120 sec and declining at longer fixed intervals. The relation between water intake and interreinforcement interval was complexly related to two drinking measures: (1) the probability of drinking following a pellet and (2) the amount drunk per bout. Drinking rate was also bitonically related to interreinforcement interval.Key words: schedule-induced polydipsia, reinforcement frequency, rhesus monkey adjunctive behavior, fixed-interval schedule, Falk (1961) originally demonstrated that rats, when food deprived and permitted to lever press for small pellets of food delivered at spaced intervals, drink excessive amounts of water in each postpellet period. This pattern of drinking has been labelled "scheduleinduced polydipsia". One factor determining the magnitude of schedule-induced polydipsia is the length of time between pellet deliveries, the interreinforcement interval. Falk (1966) reported that as the interreinforcement interval of fixed-interval schedules increased from 2 to 180 sec, total water intake also increased. When the interreinforcement interval increased to 300 sec, total fluid consumption decreased sharply, yielding an inverse-U or bitonic drinking function over the total range. Subsequent investigators (Bond, 1973;Flory, 1971;Hawkins, Schrot, Githens, and Everett, 1972;Keehn and Colotla, 1971) Deadwyler (1965) recorded the relations between a number of drinking measures and interreinforcement interval in rats exposed to response-independent foodreinforcement schedules. They reported the following effects as the interreinforcement interval increased from 30 to 480 sec: (1) the percentage of pellets followed by drinking was bitonically related to interreinforcement interval reaching peaks at intervals between 120 and 240 sec; (2) both the postpellet drink latency and drink duration increased monotonically throughout the interreinforcement interval range; and (3) the rate of licking during a drinking bout decreased monotonically throughout the interreinforcement interval range. However, they neglected to report session water-intake values to which these behiaviors might be related. More recently, Keehn and Colotla (1971) reported water intake data that could be compared with the frequency of postpellet drinking but with no other quantitative measure. They demonstrated in rats that the amount of water ingested per pellet and the probability of drinking after a pellet were bitonically related to interreinforcement interval, and both reached peak values at an interreinforcement interval of 60 sec.The present study presents systematic data on the relation between interreinforcement 257 1976, 26,[257][258][259][260][261][262][263][264][265][266][267] NUMBER 2 (SEPTEMBER)