Thirty-nine ambulant children (22 with hemiplegia, 17 with diplegia) with spastic cerebral palsy receiving isolated gastrocnemius muscle injection with botulinum toxin A were studied prospectively. The children had a mean age of 6 years (range 3 to 13 years). Measurement of gastrocnemius muscle length was used to estimate the dynamic component of each child's spasticity and to quantify the response. There was a strong correlation between the dynamic component of spasticity before injection and the corresponding magnitude of the response after injection. Children undergoing repeated injections showed similar correlations. A strong correlation was found between the duration of response and the dynamic component. Children with hemiplegia showed twice the duration for a given dynamic component compared with those with diplegia when injected with the same total dose per unit body weight. Long-term lengthening did not occur for the cohort, although some patients showed a response at a 12-month follow-up. By delaying shortening, the injections may have a role in delaying the need for surgery. Injections were well tolerated with few side effects.