THE AIM OF THIS STUDY WAS TO COMPARE CONSTRUCTION AND STORAGE EXPLANATIONS OF THE FACILITATING EFFECT UPON RECALL OF ASSOCIATIVE RELATIONSHIPS BETWEEN WORDS EMBEDDED IN CONNECTED DISCOURSE. ONE GROUP OF 36 UNDERGRADUATE SUBJECTS WAS EXPOSED TO A PASSAGE THAT CONTAINED ASSOCIATIVELY RELATED WORD PAIRS, WHILE TWO OTHER GROUPS WERE GIVEN PASSAGES THAT CONTAINED ASSOCIATIVELY UNRELATED WORD PAIRS. THE CONTEXTS OF THE VARIOUS PASSAGES WERE IDENTICAL. THE TASK WAS TO LEARN THE PASSAGE. AFTER A SINGLE STUDY TRIAL, ALL SUBJECTS WERE GIVEN A TEST OF RECOGNITION MEMORY FOR THE CRITICAL WORDS. THE CRITICAL WORDS WERE TAKEN OUT OF CONTEXT AND RANDOMIZED WITHIN A SERIES OF FILLER ITEMS. THUS, THE OPPORTUNITY FOR THE CONSTRUCTION THAT IS LIKELY TO OCCUR DURING RECALL OF CONNECTED DISCOURSE WAS REDUCED. THE GROUP EXPOSED TO THE HIGH ASSOCIATION PASSAGE WAS FOUND TO BE SUPERIOR TO THE OTHER GROUPS ON A MEASURE OF PAIR-RECOGNITION THAT WAS CORRECTED FOR THE TENDENCY TO GIVE FALSE POSITIVES. THESE RESULTS SUGGESTED THAT ASSOCIATIVELY RELATED WORDS OCCURRING IN CONNECTED DISCOURSE ARE STORED MORE EFFICIENTLY THAN ASSOCIATELY UNRELATED WORDS. THIS REPORT APPEARED IN "STUDIES IN LANGUAGE AND LANGUAGE BEHAVIOR, PROGRESS REPORT IV," 1967, CENTER FOR RESEARCH ON LANGUAGE AND LANGUAGE BEHAVIOR, UNIVERSITY OF MICHIGAN, 220 EAST HURON STREET, ANN ARBOR, MICHIGAN 48108. (AUTHOR/JD)