The effect of nonrewarded trials which preceded any rewarded trials was studied with rats in a straight runway. Running speeds were measured for groups given either 24 nonrewarded trials preceding 24 rewarded (24N-24R), 24 nonrewarded trials interpersed among 24 rewarded (50%), 24 rewarded trials (24R), or 48 rewarded trials (48R). At the end of acquisition, the mean speeds were ordered as: 50%>48R>24R>24N-24R, with Group 24N-24R showing marked forgetting on the first trial of each day after reward was introduced. During extinction the ordering was: 50%>48R>24R>24N-24r on the 1st trial of each extinction day, but 50%>24N-24R>48R = 24R on the last trial of each extinction day.