I. 12 albino rats, equally divided as to sex, were trained to get food from a food box on either the right or left side of an open field. When this habit was firmly established the reward was shifted to the opposite food box, and training was continued until the rats were consistently running to the second goal. After training had proceeded to this stage the animals were given shock in the starting box until they refused to cross the grid. Only 1 run a day was given to each animal. If the animal, after shock, proceeded to the reward box of Habit 1, the run was termed regressive. If the animal ran to the reward box of Habit 2, the run was termed perseverative. The typical regressive run was made at a much higher speed, with an avoidance of contact with the walls of the apparatus, and with hesitation. The typical perseverative run was slow; the animal hesitated and retraced its path, staying in contact with the wall. The author believes that further experimentation upon motivation is needed for a satisfactory explanation of regression. II. Rats were trained to secure food from a food box on either the right or left side of an open field. The influences of motivation were studied and the following conclusions were reached: When animals are given an electrical shock after being trained in a situation offering only one mode of response they will fixate on that mode of response for a relatively long time after the response has become inappropriate. When animals are trained in 2 different modes of response, and are satiated after reaching a criterion of master on the later-learned habit, they will revert, for the time of the satiation, to the earlier learned habit. When animals are trained in a mode of response under conditions of partial satiation, and are later trained in another mode of response under conditions of full motivation, electrical shock administered during the performance of the later-learned habit will be inefficacious in producing regression. When rats are trained under conditions of full motivation in one mode of response, and are then trained in another mode of response under conditions of lowered motivation, they will fail to eliminate the 1st mode of response, providing the reward may be gained by performing the 2 responses successively.