When reinforcements were presented after or between nonreinforcements, an induction phenomenon, i.e. the decrement of behavior variability and facilitation of running response in rats were found in the previous experiments with a semicircular maze. The present experiment was designed to examine the processes of growth and decay of the induction phenomenon by varing the number of extinction trials and the time elapsed after extinction.<br>After 140 reinforced trials (20 trials per day for 7 days), 6 groups of 12 rats were given 0, 10, 20, 40, 60 and 60 extinction trials, respectively. Immediately after extinction, they were run 20 reinforced trials (reconditioning period) except the last group which was given the same procedure after a 23hr rest period.<br>The results obtained during the reconditioning period were:<br>1) With increasing number of extinction trials, decrease of behavior variability (in four measures) and facilitation of running time were found in 5 groups which were given the reconditioning period immediately after extinction. The analysis of variance of the data showed that the group difference was significant in each of above 5 measures. The t-test with Ryan's correction yielded significant differences between the 0 trial group and the other 4 groups on variability and running time scores. Significant differences were also found between 10 trials group and 20, 40 and 60 trials on variability scores. However, there was no significant difference between the 40 trials group and the 60 trials group on all measures.<br>2) When the present data were compared with those of the previous studies in which nonreinforcements were given partially during daily sessions (i.e. partial reinforcement, successive discrimination, and repetitive conditioning-extinction procedures), the present experiment (i.e. continuous nonreinforcements) was found to develop the greatest amount of induction. This might be caused by the difference of number of reinforcements which was given before the first nonreinforcement: that is, the more reinforcement, the more induction.<br>3) The last group tested 23hr after 60 extinction trials showed reduction of the induction during this period since all scores recovered to the level before extinction. We had another data on two groups which were tested 23hr after 120 and 300 extinction trials (Fujita and Ishihara, 1963; Ishihara, unpublished data), so we could compare these three data in order to examine the effect of the number of extinction trials upon induction after a 23hr rest period. It was found that amount of induction was positively related with the number of extinction trials. As shown in 1), however, when reconditioning was given immediately after extinction, there was no difference in induction between the groups which received extinction trials more than 40 trials. Therefore, it is suggested that, although the amount of induction was almost identical because of the ceiling effect immediately after extinction, the differential effect was revealed after a 23hr intertrial interval.