Purpose: This study explored the effects of 2 different training structures on the implicit acquisition of a sequence in a serial reaction time (SRT) task in children with and without specific language impairment (SLI). Method: All of the children underwent 3 training sessions, followed by a retention session 2 weeks after the last session. In the massed-training condition, the 3 training sessions were in immediate succession on 1 day, whereas in the distributed-training condition, the 3 training sessions were spread over a 1-week period in an expanding schedule format. Results: Statistical analyses showed that the children with normal language were unaffected by the training conditions, performing the SRT task similarly in both training conditions. The children with SLI, however, were affected by the training structure, performing the SRT task better when the training sessions were spaced over time rather than clustered on 1 day. Conclusion: This study demonstrated that although intensive training does not increase learning in children with SLI, distributing training sessions over time does increase learning. The implications of these results on the learning abilities of children with SLI are discussed, as are the mechanisms involved in massed versus distributed learning.