The co-use of cocaine and heroin is relatively common, with a growing clinical and preclinical literature dedicated to investigating the factors underlying the phenomenon. Specifically, several studies have compared the reinforcing effects of the coadministration of cocaine and heroin, referred to commonly as 'speedball', to either drug alone. The present study assessed whether addition of heroin to a wide range of cocaine doses produces reinforcing effects greater than cocaine alone using both a progressive ratio (PR) schedule and a choice procedure. Patterns of coadministration of cocaine and heroin offered simultaneously were also assessed using double-lumen cannulas. Under the PR schedule, speedball combinations across a range of doses (0.38-3.0 mg/kg/inf cocaine+1.5-48 microg/kg/inf heroin) did not support higher break points than cocaine alone. When cocaine and heroin were made available concurrently (ie on two separate levers), rats self-administered cocaine exclusively. Using a choice procedure, however, a preference was demonstrated for some speedball combinations (eg 0.18 mg/kg/inf cocaine+50 microg/kg/inf heroin; 0.38 mg/kg/inf cocaine+50 microg/kg/inf heroin) over cocaine alone (0.75 mg/kg/inf). So while results obtained using the PR schedule do not support the hypothesis that speedball combinations are more reinforcing than cocaine alone, data from the choice procedure do support this hypothesis. These apparently discrepant results demonstrate that these models are measuring different aspects of drug reinforcement, and suggest that choice procedures in rats provide a useful tool to study speedball self-administration.