This study tests the hypothesis that components of play such as immagination contribute to cognitive development. Twenty-four middle-class children attending a university-affiliated preschool were observed for 20 one-minute play observations during the Spring Semesters of 1978 and 1979. Play was coded using categories for both social (solitary, onlooker, parallel, and interactive) and cognitive (functional, constructive, and imaginative) components. Imaginative play was further scored for number and type of transformations (person, object, situations) and for thematic content (sociodramatic or fantasy). During each semester, children were administered the Peabody Picture Vocabulary Test (Form B) and the Ravens Progressive Matrices tests of intelligence, and a battery of classification (dichotomous sort, cross-classification, and class inclusion) and conservation (liquid substance, number, length, area, and weight) tasks. Pairing of play and cognitive test change scores from year one to year two revealed that, generally more children increased in cognitive ability without increasing in imaginative play than did those who increased in dramatic play without increasing in cognitive ability. Results suggest that directional influence between these two variables moves from cognitive ability to imaginative play and not vice versa. (Author/RH)