The research reported in this paper is the result of a project to develop a preschool mathematics curriculum for economically deprived children in Hawaii. The developed materials were tested with preschool children in Head Start classes and a preliminary assessment of the curriculum was made. An attempt was made to evaluate the curriculum which was developed by comparing gains from pretest to posttest in several measures of mathematical knowledge and understanding for classes that had been trained for twenty minutes per day with the new curriculum versus classes which received no training. Tasks presented to the children in the experimental groups related to number and counting, geometry, dimension, prearithmetic operations, and symbols. The Geometric Design, Arithmetic, and Block Design subtests of the WPPSI; an experimental form of the Head Start Arithmetic Test; and two Piagetian conservation tasks were administered to the children in three experimental classes and two comparison classes. Using an analysis of covariance, results on the Geometric Design (p=.001) and Block Design (p=.05) subtests were found to be significantly in favor of the experimental group when the total experimental group was compared to the total comparison group. The amount of change between pre- and posttesting with the Head Start Arithmetic Test was found to be significantly greater (p=.02) for the experimental than for the comparison group. (FL)