The problem raised in the present paper deals with the role of reciprocity in complex structures. By definition, a three-entity situation in which all six relations are negative is unbalanced, yet all the relations are re ciprocated. A similar three-entity structure in which these six relations are positive is balanced and completely reciprocated, as well. The question which is raised is whether reciprocity is an independent source of cognitive bias. From the results of DeSoto and Kuethe 1959 it would appear that with respect to the liking relationship reciprocity may indeed constitute a source of cognitive bias. Subjects in their experiment estimated the probability of reciprocation of the liking relation to be .74. In a further experiment by DeSoto 1960, Ss took longer to learn a hypothetical social structure characterized by absence of reciprocity than one in which all relations were reciprocated. It was not clear from the results of either experiment, however, if these learning effects were due to reciprocity alone or to balance, since the structures employed by these researchers differed from one another not only in terms of reciprocity but in terms of structural balance as well.