Most medications prescribed for attention-deficit-hyperactivity disorder are psychomotor stimulants with reinforcing effects in laboratory animals (eg methylphenidate). The present studies were conducted to evaluate the reinforcing effects of the recently approved medication atomoxetine in monkeys trained to 'choose' between automated deliveries of either an i.v. injection or food. Rhesus monkeys were trained to lever-press under concurrent schedules of reinforcement; responses on one lever resulted in an injection of either saline or drug, and responses on the alternative lever resulted in food delivery. Data were collected on four measures: response rates, percentage of total responses occurring on the injection-lever (% ILR), number of injections earned, and number of food pellets earned. Dose-effect functions were determined for cocaine (0.003-0.3 mg/kg/inj), methylphenidate (0.003-0.1 mg/kg/inj), amphetamine (0.003-0.1 mg/kg/inj), atomoxetine (0.01-0.3 mg/kg/inj), and desipramine (0.03-1.0 mg/kg/inj) using a double alternation schedule of saline and drug availability. Results indicate that the distribution of behavior changed according to the drug and dose available for self-injection. Saline availability was typically associated with high rates of food-maintained responding. The % ILR increased from 3+/-1% when saline was available to >90% when >0.03 mg/kg/inj of cocaine, methylphenidate or d-amphetamine was available. However, no dose of atomoxetine or desipramine maintained self-administration behavior on the injection-lever. The number of food pellets earned per session decreased as the dose of each drug increased, indicative of behavioral activity with all five drugs. The reinforcing effects of cocaine, methylphenidate, and d-amphetamine in these studies are consistent with previous findings in nonhuman primates and with their documented abuse liability. The absence of reinforcing effects of atomoxetine support the view that, like desipramine, it has no evident abuse potential.