A recent therapeutic intervention, constraint-induced movement therapy (CIMT), has been shown to improve movement efficiency and quality of movement in the involved hand of children with hemiplegic cerebral palsy (CP). In the present study, we investigate the long-term effects of CIMT and the effect of a second course on involved limb function using an ABABA design. Eight children with mild to moderate hemiplegic CP (six males, two females; mean age 8y 7mo [SD 2y 6mo]; range 5-11y), who had received a CI therapy intervention 12 months before this study, participated in a second intervention. In both interventions, the children wore a sling on their non-involved upper extremity for 6 hours per day during 10 out of 12 consecutive days and were engaged in play and functional activities that provided structured practice using the involved upper extremity. The results indicated initial improvements in movement efficiency, as measured by the Jebsen-Taylor Test of Hand Function, the Speed and Dexterity subtest (no. 8) of the Bruininks-Oseretsky Test of Motor Proficiency, and caregivers' perceptions of amount of use and quality of movement of the involved limb, were retained 12 months after the first intervention. The second intervention resulted in further improvement on these measures. Results indicate that intensive practice associated with CIMT may be retained long term, and that continued improvements may occur after a second intervention dose.