The drug self-administration reinstatement procedure provides an important animal model of relapse. While the procedure is widely used, there has been little investigation of the basic processes involved. This experiment determined the specificity of reinstatement by examining reinstatement of food seeking by D-amphetamine. During training, 24 rats pressed levers for food. Eight rats received 3.0 mg/kg D-amphetamine before and saline after sessions. Eight rats received saline before and after sessions. The final eight rats received saline before and 3.0 mg/kg D-amphetamine after sessions. All rats then experienced saline injections and extinction. During a reinstatement test, all rats received 3.0 mg/kg D-amphetamine. D-Amphetamine significantly increased lever pressing for rats with prior exposure to amphetamine as a predictive cue for food (pre-session) and for rats with no prior exposure to amphetamine. The effect was larger for rats with pre-session exposure to amphetamine than for rats with no previous exposure. Rats with exposure to amphetamine but not as a predictive cue for food (after sessions) did not show significant reinstatement of lever pressing. Therefore, the reinstating effects of amphetamine are not restricted to behavior previously maintained by amphetamine self-administration. In animal models of relapse, reinstatement of drug seeking could be due, in part, to discriminative and direct effects of self-administered drug.