The generation of adjunctive behavior was studied in rhesus monkeys responding under fixed-interval (FI) schedules of food pellet delivery or intravenous (iv) infusions of cocaine or midazolam. The FI schedule value ranged from 1.5 to 120 min. Each response on a second, concurrently available lever resulted in an infusion of saline. The number of saline infusions received was a bitonic function (inverted U) of the interval of reinforcer (food or drug) delivery. This bitonic function is characteristic of adjunctive behavior and strongly suggests that the responding that results in the infusion of saline was an adjunctive behavior. Furthermore, these results establish that intermittent schedules of drug delivery can generate adjunctive behavior. Increases in the dose of drug self-administered shifted the bitonic function of saline infusions to the right. This shift in the function of adjunctive behavior is consistent with the hypothesis that the generation of adjunctive behavior is dependent upon both the magnitude and the intermittent delivery of the reinforcer.