The persistence and generality of the contextual interference (CI) effect was tested using a rapid sequential aiming task. Participants (N=48) practiced three movement patterns for three blocks of 18 trials under a blocked (BL) or random (RA) schedule. Movement patterns were displayed and KR provided throughout practice and testing. A 24-hr delayed knowledge of results (KR) retention test included three blocks of 18 trials, followed by a transfer test of a single new movement pattern. In contradiction of the CI effect, RA practice provided an advantage for RA retention only, not BL retention. Furthermore, group differences at transfer were not persistent. Hence, RA practice does not necessarily provide enhanced, context-free learning, but it is essential training for task switching. Overall the findings reveal limited persistence and generality of the CI effect.