Objective and subjective results from nine experiments are presented. The conclusions concerning the existence of mediate association, which the author contends must naturally vary with the definition adopted, are as follows: (1) If the concept be defined as the tendency for two items to recur together in recall, because they have previously been associated with a common item, the existence of this kind of mediate association is proven by the data of these experiments. (2) If the term be defined as a tendency for one item to arouse the other in recall when the common item is not present at the time, the concept is false. (3) The third possible definition assumes that one of the two items will arouse the other when the common item is perceptionally or ideationally present at the time of recall. The present data support the truth of this concept. (PsycINFO Database Record (c) 2006 APA, all rights reserved)