A general working model of cognitive development assumes that there are sets of orthogonal cognitive abilities, which remain fairly stable after age 7. This paper examines the long term predictive and diagnostic value of assessing specific cognitive abilities among preschool children. This model by empirical studies was defendable on the grounds that the methodology of group empirical studies tended to prejudice results in favor of a general cognitive ability model. The assessment techniques used in this study draw heavily from a perceptual survey rating scale developed by Kephart for primary grades. Tests were administered to 74 middle class nursery school, 4- and 5-year-olds. The tests consisted of three visual pursuit tasks; measures of convergence, refixation ability, and power; and power; and the Draw-A-Circle task. In addition, the preschoolers were administered the Stanford-Binet, a specially developed preschool achievement test, and a measure of impulsivity control. Data was factor analyzed. Several problems identified were lack of observer agreement, unclearness as to what several of the tests were actually measuring, and scoring difficulties on the Draw-A-Circle. (MH)