The efficacy of d-amphetamine to support a selective bilateral intra-accumbens self-administration response was examined. Bilateral intra-accumbens infusions of d-amphetamine were made contingent upon the acquisition of a lever-pressing response. Two identical levers were available within the operant chamber. Depression of the drug lever resulted in the intra-accumbens delivery of 1 microgram d-amphetamine; responses upon the second, control lever were recorded but had no programmed consequences. Animals were not 'primed' with non-contingent infusions of d-amphetamine at any time during these experiments. Nonetheless, animals readily acquired a selective response upon the drug lever. Removal of the d-amphetamine moiety from the infusate resulted in a large decline in responding, and the abolition of the selectivity of the response for the drug lever. Adulteration of the infusate with either the D1 dopamine receptor antagonist SCH-23390 or the D2 dopamine receptor antagonist sulpiride enhanced the rate of response selectively upon the drug lever. Reductions in the dose of d-amphetamine also increased the rate of response. The effect of co-adulteration of the infusate with both SCH-23390 and sulpiride together was purely additive. The implications of these data for the methodology of intracranial drug self-administration, and the relationship between D1 and D2 dopamine receptors within the nucleus accumbens are discussed.