Three-hour sessions of differential reinforcement of low-rate behavior (DRL 45-s) in rats were used to investigate alprazolam, caffeine, and their interactions at the onset, peak and disappearance of serum alprazolam, while caffeine level remained constant. The dose-response curve (DRC) method of Pöch permitted an extensive evaluation of the kind (additivity or independence) of interactions occurring in combined drug effects. The alprazolam and caffeine DRCs were used to derive theoretical additive and independent relations, and the observed combined effects compared to these functions. Behavior-time profiles of the combined effects were similarly compared. Serum alprazolam and caffeine concentrations correlated with their respective behavior-time profiles. No acute tolerance was observed either for the individual drugs or their combinations. Alprazolam was more potent than caffeine in disrupting DRL behavior. Because alprazolam is much shorter-lived (t1/2 = 32 min) than caffeine (t1/2 = 3 h) in rats, potency ratios between alprazolam and caffeine changed across session time (from 123 to 4), which determined the expression of the combined effects. Although the combined effects were not distinguishable in terms of additivity or independence in both the DRCs and in the behavior-time profiles, they showed neither synergism nor antagonism. The pharmacokinetics of alprazolam were not altered by caffeine, but those of caffeine were affected by alprazolam.