Rats were trained to discriminate between two compound stimulus arrays that included exteroceptive (ES) and interoceptive (IS) stimulus components. The ES components were illumination and tactile cues, and the IS components were produced by d-amphetamine (A) or Na pentobarbital (P) treatment (drug versus drug discrimination). The relative ES-IS strength was then determined by testing with "mismatched" stimulus arrays consisting of combinations of ES and IS components that had not been presented simultaneously during training. Additional tests were done with ES only (no drug treatments). At training doses of 0.8 mg/kg A versus 10 mg/kg P, the ES were less salient than, but did share stimulus control with, the stronger IS components. In a second group, trained with lower doses (0.5 mg/kg A versus 4 mg/kg P), the ES were much more salient than the IS, although again, both types of stimuli did acquire some control. This dose-related sharing of stimulus control between ES and IS components is similar to relationships among components of compound arrays consisting entirely of ES, and thus further illustrates similarity between drug-produced and exteroceptive stimuli.