24 7-yr.-old children and 24 university-age adults practiced a ballistic aiming task under either low contextual interference (blocked practice) or high contextual interference (random practice). All subjects performed 90 acquisition trials, followed by 20 transfer trials and 15 retention trials. Similar to previous findings, the adults performed the acquisition trials better under blocked than under random conditions, yet performed the retention and transfer tests better after random than blocked practice. No differences in acquisition were found between blocked and random practice conditions for the children; nevertheless, the random group performed the retention and transfer tests better than the blocked group. The results are discussed in relation to applied and theoretical issues of contextual interference.