Research related to determining how procedural variables can alter dose-effect functions for cocaine self-administration is limited. Toward clarifying the role of procedural variables, responding was maintained in rats under either variable-interval (VI) or fixed-ratio (FR) schedules of cocaine infusion. In addition to free-operant FR schedules, discrete-trial FR schedules were evaluated. The dose-effect functions were obtained by either substituting a dose for the usual daily dose, instituting a particular dose for several sessions, or making all doses available within a session. Dose-effect functions for response rate (or number of trials with infusions for the discrete-trial FR) were bitonic for the VI and discrete-trial FR schedules but tended to be strictly decreasing for the free-operant FR schedules. Responding was maintained under FR schedules by a low dose (0.083 mg/infusion) if the dose was substituted for a higher daily dose but not when made available daily. Rate of response was higher under ratio schedules at 0.17 mg/infusion when this dose occurred within the context of other higher doses within a session than when the dose was simply substituted for a higher daily dose. These data indicate that procedural variables can alter dose-response curves for cocaine self-administration.