Young (2) found the constancy of the IQ to be influenced by the length of interval between tests, types of tests used, IQ level, sex, and chrono logical age. In contrast, Bonney (12) obtained correlations ranging from +.71 to +.93 between IQ ratings separated by an interval of one year. His samples, however, were small. Goldin and Rothschild (30) found correlations of +.80 or higher, for children of Italian parentage, between Henmon-Nelson scores originally obtained in the fourth grade and retests with the same test at the sixth-and eighth-grade levels; but the corre lation was only +.46 for Pintner-Cunningham scores obtained in the first grade and Henmon-Nelson scores obtained in the fourth grade.Hildreth (35) completed her analysis of 149 train and locomotive drawings selected from a continuous series of more than two thousand drawings made by one child between the ages of two and eleven years. A smooth and consistent growth curve, with no sharp breaks or regres sions during the nine-year period, was obtained.Summary statements pertinent to future longitudinal studies were made by Hamalainen (33) on pupil growth in the elementary school, and by Jersild and Fehlman (40) on developmental objectives.