Studies of conceptually mediated behavior in human beings are often concerned with the effects of variations in previously established concepts on behavior in the solution of problems. This is true, for example, of the work of Maier (1930) with practical problems; of Luchins (1942) on water-jar problems; and of Maltzman and his collaborators (Malzman, Eisman, Brooks, & Smith, 1956;Maltzman & Morrisett, 1952, 1953a v , 1953b on anagrams; to mention a few well-known investigations. Usually, in such studies, the concepts employed by 5 are assumed to have a history of previous establishment, and are considered to be available to 5 at the time the problem is set. In a different category can be placed research on the learning (or formation) of concepts, such as that of Hull (1920), Smoke (1932), Reed (1950), and Heidbreder (1947, which need not be reviewed here. The concepts acquired in the course of the experiment are usually not further "used," as in the solution of a problem, but are