to the point of one correct recitation, the subject was required to study it further and to recite it daily for a period of one week. The passage proved to be very difficult; but in learning it the subject formed very strong associations between the consecutive words or groups of words. It would seem at first thought that the distributed practise on this long difficult passage should afford ample opportunity for the formation of remote forward associations. Relearning the F- and S-Lists.-After the prose passage had been relearned and recited daily for a period of one week, the subject relearned these lists, using the same method as before. After completing, each subject was asked a number of questions about the experiment. All thought that most of the words in the 25 lists had been taken from the prose passage, but they did not notice the method used in constructing them, and no subject had any insight into the problem being investigated. All subjects stated that various parts of the prose passage would intrude themselves, or were recalled from time to time, while they were relearning the lists. This tendency was more pronounced in some lists than in others. Almost all of the subjects thought that this recall of parts of the prose passage facilitated the relearning of the lists, by making them more familiar, and giving them more meaning. The objective results, however, showed that these opinions were not reliable. The S-lists were made up of words arranged in the backward direction, when the words in these lists are compared with the arrangement of the same words in the prose passage. They are composed of words which were originally separated from each other in the prose passage by 21 intervening words. The F-lists, on the other hand, are made up of words arranged in the forward direction, and separated from each other by only I intervening word, when they are compared with the order of the words in the prose passage. Now when the prose passage is learned, it can be assumed that the principle of remote association should establish much stronger associations between every 2d word in the forward direction (F-lists) than between every 22d word in the backward direction (S-lists). It can be assumed in fact that the strength of the associations between every 22d word in the backward direction (S-lists) is practically zero. If The S-lists represent a zero base of learning in the present experiment, and the presence or absence of remote forward associations may therefore be determined on the basis of a comparison between the efficiency of learning the S-lists and the F-lists, before and after learning the prose passage. the efficiency of learning the F-lists (compared with the S-lists) improves after learning the prose passage, a conclusion in favor of remote forward association is justified. If learning the prose passage reduces the efficiency of learning the F-lists (compared with the S-lists), then it follows that learning the prose passage inhibits the formation of remote forward associations. Results Influence of Learning Difficult Prose Passage on Remote Forward Associations; Individual Results.-The individual results for the 17 subjects are summarized in Tables IA, B and C. The general method of treating the data may be illustrated by the results of the first subject (Table IA). The first column in the table shows that before learning the prose passage the 11 F-lists required an average learning time of about 86 sec, and the II S-lists an average learning time of about 97 sec. The superiority of the F-lists over the S-lists therefore is about II sec. This difference is expressed in per cent. by dividing 11 by the average of 86 and 97. After learning the prose passage, the same F-lists required an average learning time of 55 sec, and the same S-lists an average of 63 sec. The difference is now + 8 sec, in favor of the F-lists. If the figures in the last two lines of the table are positive, the conclusion is in favor of remote forward association. If these figures are negative, the results indicate that learning the prose passage has resulted in a condition which inhibits the formation of remote forward associations. The percentage difference given in the last row of the table is more significant than the difference (d-c) - (b-a), because of the shift in the base of comparison after learning the prose passage. TABLE IA AND IB INFLUENCE OF LEARNING DIFFICULT PROSE PASSAGE ON REMOTE FORWARD ASSOCIATIONS; INDIVIDUAL RESULTS 2 (a) F-Lists. 86.3 (b) S-Lists 81.0 53.9 97.7 84.0 61.5 +11.4 3.0 + 7.6 +12.4 +3.6 +13.2 7.0 +17.6 0.0 -10.8-13.0 + 7.8 -15.7 +19.2 0.0 -13.0 17.0 +12.7 59.0 54.1 51.0 +12.0+ 3.2 +22.6+ 6.1 + 5.0 +10.3 After Learning Passage: (c) F-Lists. 55.2 47.0 34.2 30.0 40.6 38.0 59.3 57.0 44.5 42.0 28.2 26.0 62.9 48.0 + 7.7 +1.0 +13.0 + 2.0 40.1 35.0 63.2 45.0 65.5 48.0 44.5 46.0 32.0 36.0 + 5.9 + 5.0 +22.6 + 7.0 + 6.2 9.0 0.0 + 4.0+ 3.8 +10.0 +15.9 +15.4 +43.5 +16.9 + 9.9 +17.1 0.0 + 9.1 +12.6 +32.3 65.2 58.0 71.6 68.0 125.0 100.0 70.0 81.3 64.8 65.0 91.0 81.0 107.2 86.0 103.2 99.0 0.4 7.0 +19.4 +13.0-17.3-39.0+ 3.2 +29.0 0.6 +11.4 +23.9 +17.4 -14.9-37.0+ 3.1 +34.3 75.0 57.7 58.0 94.3 +13.0 +12.0 +14.8 87.0 75.9 71.0 +18.2 +13.0 +14.8 +27.2 +20.2 After Learning Passage: (c) F-Lists (d) S-Lists d-c. (Per cent.). Effect of Learning Passage, or (d-c) — (b-a) (Per cent.). + 5.4 +7.0 -15.9 16.0 +27.0 +56.0 + 6.2 +14.4 +27.5-15.8-24.6 +31.4 +68.89.2-21.7 +14.0 +40.0 8.3 6.4 - 20.0 6.5 +1.0 0.8 +19.8 -29.9 TABLE IC INFLUENCE OF LEARNING DIFFICULT PROSE PASSAGE ON REMOTE FORWARD ASSOCIATIONS; INDIVIDUAL RESULTS These tables show that there is considerable variability in the results of the individual subjects. Some give results which are definitely in favor of remote forward association, while the results of others are just as definitely opposed to this theory. If the values for (d-c) (b-a) are considered, half of the subjects give results in favor of remote forward association, and half give the opposite results. If the percentage differences are considered, the results are very slightly in favor of the theory. There is very little correlation between the type of the results and the nature of the individual subjects. Thus the most naive subjects are nos. 3, 6, 7, 8, 10, 13, 16, and 17, 4 of whom give results for and 4 results against the principle of remote forward association. The subjects best versed in psychology are nos. 9 and 14. No. 9 gives results in favor, and no. 14 results against the theory. The most conscientious subjects are nos. 6, 7, 13, and 17. Two of these give results for, and 2 against. Five of the subjects could be classified as slightly "peculiar" (nos. 1, 5, 11, 15, and 16). Two of these give results for, and 3 against. The most distinctly "abnormal" subject in the group is no. 15, who gives results in favor of the theory. The uncontrolled factors in the situation and the use of different learning methods seem to afford a plausible explanation for the variability of the results. Influence of Learning Difficult Prose Passage on Remote Forward Associations; Combined Results of 17 Subjects.-All of the gross time scores for the 17 subjects are arbitrarily combined in Table II. The general method of presentation is already familiar. The table shows that if the values for (d-c) - (b-a) are considered, the results are slightly opposed to remote forward association. If the percentage differences are considered, the results are very slightly in favor of the theory. The results as a whole seem to indicate that, with the procedure used, the chances of obtaining a result in either direction are practically equal. Although the variability of the individual results described above is significant from a practical point of view, the results themselves do not exclude the possibility that there may be a |