Every one of 40 subjects learned 12 lists of 10 non-sense syllables and 12 test-lists made by replacing the even-numbered syllables of each list with the even-numbered syllables of another one of the original lists, so that each syllable retained its original position in the list. Control-lists were made by taking the first syllable from the first original list, the second syllable from the second, and so on. Hull's exposure apparatus was used; exposure was 1.5 sec. for each syllable. The number of repetitions required for one correct recitation was obtained for learning test- and control-lists immediately after learning the original lists and a week afterwards. For immediate relearning the test-lists show 4.9% more savings over the original lists than do the control-lists; after a week the savings amount to 33.4%, statistically a highly reliable difference. "The functioning of remote associative tendencies is inhibited in proportion to the strength of the original learning and is facilitated, within certain limits, in proportion to the period of forgetting." (PsycINFO Database Record (c) 2006 APA, all rights reserved)