The purpose was to examine the effects of three different demonstrations by a model on acquisition and retention of a sequential gross movement task. The second purpose was to examine the relationship between reversal processing of visual information about skills and coding of skill information. Thirty undergraduates (15 men and 15 women) were assigned into one of three conditions, Objective condition which demonstrated the task with the model facing the subject. Looking-glass condition in which the skill was demonstrated with the model facing the subject who viewed the performance opposite the right and left directions in executing the task, and the Subjective condition in which the subject observed the model from the rear. Number of immediate recall tests required to accomplish the sequential movements completely and the sum of the performance points for reproduced movements at each delayed recall test (1 day, 7 days, and 5 mo. after the immediate recall test) were employed. Analysis indicated the Subjective condition produced a significantly greater modeling effect in immediate recall of the movements than the Looking-glass condition. Retention of the acquired skills was almost equal under the three conditions.