In a nine-week double-blind crossover comparison of dextroamphetamine, levoamphetamine (Cydril), and placebo with 11 hyperkinetic children, effects were assessed by an established teacher rating scale, a parent rating scale, and a new tool, weekly quantification of parent-selected target symptoms. Both active drugs were significantly more effective than placebo. Dextroamphetamine seemed consistently superior to levoamphetamine, though not to a significant degree (on this size sample). Levoamphetamine seemed "slower starting," requiring three weeks to show significant benefit on target symptoms, whereas dextroamphetamine showed nearly its maximum benefit the first week. Levoamphetamine seemed better for hyperactivity and aggressiveness than for inattentiveness, whereas dextroamphetamine seemed equally beneficial for all three. These data are consistent with the possibility that therapeutic effects of amphetamine on hyperkinetic children are mediated, at least in some, by dopaminergic systems.