This study is concerned with misarticulated speech sounds of children and the phonetic realization of these sounds. The articulation errors of 384 standard-English-speaking school children were analyzed in speech samples obtained by the National Speech and Hearing Survey and were samples of both free speech and of performance on the Goldman-Fristoe Test of Articulation. Error rates and incidence of various types of errors (omissions, distortions, and substitutions) by grade level were noted. The major thrust of the investigation, however, was an attempt to explain the substitution errors in the sample using the constructs of distinctive feature theory and markedness theory. The hypothesis presented in and confirmed by this study is that when substitution errors occur, less complex phonemes will be substituted for more complex ones. The data analyzed in the present study suggest that substitutions are governed in part by a tendency toward ease of articulation with constraints imposed upon substitutions by a tendency to maximize perceptual distinctions in the speech output. The phonetic features themselves were found to be of varying degrees of stability. Tables and charts are included. (Author/RL)