Rats trained to lever-press for electrical stimulation of the nucleus accumbens, lateral hypothalamus, or ventral tegmental area, were tested with a range of stimulation frequencies to assess the effects of naltrexone (2.5, 5.0, 10.0, 20.0 mg/kg, i.p.) during sessions beginning 15 or 45 min after injection. Naltrexone, when effective, shifted the rate-frequency functions to the right; the magnitude of the effect depended on site of stimulation and on the delay after injection. The greatest effect was observed with stimulation of the nucleus accumbens and the least with stimulation of the ventral tegmental area. There was a greater attenuation of responding during the late test sessions than during the early ones. The time course of naltrexone's effect on brain stimulation reward was determined for the highest dose by measuring a rat's rate of responding over a 3 h period in sessions with immediate access (5-min delay) or delayed access (45-min delay) to stimulation. The greatest decreases in responding were observed 45, 65, and 85 min after injection and the delay in access made little difference. The fact that the drug was more effective 45 min after injection explains some of the inconsistencies in the literature; the fact that its effectiveness was independent of early exposure to stimulation would suggest pharmacological rather than experiential factors as the explanation of the delayed effectiveness.