ACCORDING to Deutsch's theory1, the Olds effect2 of electrical self-stimulation is produced as a result of the concomitant simulation of both drive and reward pathways. If two such systems, one mediating drive and one reward, are indeed involved, the threshold stimulus intensities required to activate each might differ. If the drive system has a lower threshold than the reward system, then electrical stimulation of lesser intensity than that necessary to produce learning (where reward is also requisite) should prolong a habit which has already been learned, irrespective of the timing of the stimulation with respect to the performance of the habit. That is, the stimulation should prolong, say, bar-pressing, even if its administration is non-contingent on the act of pressing. If, on the other hand, such non-contingent stimulation is administered at an intensity above the threshold intensity necessary to produce learning, then a habit which the animal has already learned for intracranial stimulation should become weaker, since the animal will also be accidentally rewarded for other acts and so come to repeat these instead. Such learning of alternative habits should not occur when the non-contingent stimulation is of sub-threshold intensity for learning, and so there should be less interference with the original habit.