To assess the effects of practice and dynamic instruction on changes in speed and accuracy during acquisition of simulated laparoscopic surgical skills.Fourteen PGY-1 general surgery residents were randomly assigned to 1 of 2 experimental conditions (n = 7 per group), either practice only or practice with instruction, and required to perform 10 trials of each of 2 laparoscopic surgical skills-cannulation and object passing. Practice only subjects were given verbal instructions for each task, and corrective feedback only after trial 1. Practice with instruction subjects were treated the same, but also saw a videotaped demonstration and received dynamic feedback during and between each trial. Performance speed was recorded for each trial and number of errors was recorded for trials 8 to 10 by videotape review.Mean speed for subjects in both groups increased significantly for both tasks (p < 0.01). Practice with instruction subjects committed significantly fewer errors on object passing (p < 0.04) and were less variable in the number of errors committed during the cannulation task (p < 0.01).Practice, with or without dynamic instruction, results in significant improvement in the speed of performance of simulated laparoscopic surgical skills. The addition of dynamic instruction to simulator-based practice improves the quality and consistency of resident acquisition of laparoscopic surgical skills.