Theoretically and pragmatically the subtest questions of the Wechsler Preschool and Primary Scale of Intelligence are assumed to be ordered from least to most difficult. The purpose of this study was to test whether or not the order was appropriate for a sam le other than that used for standardization. Data consisted of test records or89 females and 62 males tested at a private school over four years. Mean Full Scale I& of the sample was 118. Calculation of proportion of the sample prtssing each item and Guttman Reproducibility Coefficients fcr each subtest suggested changes in the order of questions of the Vocabulary, Picture Completion, Similarities, and Comprehension Subtests. However, though questions were misorderd for the present sample as a result of presenting them in the order prescribed in the test manual, the misordering was not great enough to seriously affect subtest scores or the interpretation of intratest scatter.The administration and scoring of the widely used Wechsler Intelligence Scales are based on the assumption that the questions within each subtest are ordered from least to most difficult. The present study was designed to test that assumption for the Wechsler Preschool and Primary Scale of Intelligence (WPPSI).In developing his scales, Wechsler (1958), an adherent of point scale intelligence measurement, placed subtest questions in ascending order of relative difficulty according to the frequency with which each question was answered correctly by the standardization sample. But he warned that the order of difficulty could be affected by sex of the subject, his geographical location, age> and national origin. Therefore he suggested that an examiner could proceed beyond generally adopted subtest termination points as long as caution was exercised not to violate the norms of the examination. Authors studying the WPPSI's predecessors have suggested that subtest items are not in proper ascending order and have voiced clinical misgivings. Jastak (1950) studied the frequency of correct responses to WB-I subtest questions and found that deviations of item difficulty were common enough to impede the economy and ease of testing, to make the setting of reasonable limits for individual testing difficult, and to hamper accurate derivation of intratest scatters.