Successful motor skill learning requires repetitive training interrupted by rest periods. In humans, improvement occurs within and between training sessions reflecting fast and slow components of motor learning [Karni A, Meyer G, Rey-Hipolito C, Jezzard P, Adams MM, Turner R, et al. The acquisition of skilled motor performance: fast and slow experience-driven changes in primary motor cortex. Proc Natl Acad Sci USA 1998;95:861-8]. Here, these components are characterized in male and female rats using a model of skilled forelimb reaching and are compared to time scales of instrumental learning. Twenty female and 14 male adult Long-Evans rats were pre-trained to operate a motorized door (via a sensor in the opposite cage wall) to access a food pellet by tongue. Latencies between pellet removal and door opening were recorded as measures of instrumental learning. After criterion performance was achieved, skilled forelimb reaching was requested by increasing the pellet-window distance to 1.5cm. Reaching success was recorded per trial. Mean latencies decreased exponentially over sessions and no improvement within-session was found. Skill learning over eight training sessions followed an exponential course in females and a sigmoid course in males. Females acquired the skill significantly faster than males starting at higher baseline levels (P < 0.001) but reaching similar plateaus. Within-session improvement was found during the sessions 1-3 in females and 1-4 in males. Performance at the end of session 1 was not carried over to session 2. Learning curves of individual animals were highly variable. These findings confirm in rat that motor skill learning has fast and slow components. No within-session improvement is seen in instrumental learning.