An assessment was made of the strength and direction of the relationship between rated degree of familiarity for connected discourse and the extent of recall of such material. The experimental material consisted of two short passages of prose (a narrative and an argument) and of a rating scale containing all the sentences from these passages inserted randomly among other individual sentences selected from a wide range of sources. 20 Ss provided both recall scores for the passages (which were presented whole) and familiarity ratings for the sentences in the rating scale. When recall scores for the individual sentences were correlated with the mean ratings, a positive and significant value was found. The bearing of this finding on (he expectations of interference theory is discussed.