Abstract
Intertrial intervals (ITIs) oi 10 sec, 3 min., 20 min., and 24 hr. during T-maze acquisition were combined factorially with intervals of 10 sec, 3 min., and 20 min. during reversal. During initial acquisition trials spontaneous alternation was inversely related to ITI, but overall error performance was relatively unaffected by ITI. Choice performance during reversal was inversely related to reversal ITI. Running speed was inversely related to existing ITI during acquisition and reversal. Results are interpreted with an interference theory of forgetting which asserts specifically that proactive inhibition is the major source of interference. Acquisition choices are impervious to forgetting, and running speeds are not, because interfering responses are available for running but not for choosing.