This report represents the first step in a systematic study of the use of vowel and consonant elements by feeble-minded children. The purpose is to determine developmental trends in the speech of mentally retarded children and to relate data acquired in this field to the processes of development in normal infants and children and to the status of adult speech. Ultimately, information concerning the comparative rates of speech development in normal and feeble-minded children may be established and remedial measures for speech retardation may be set up. A further aim is to try out several indices of speech status on this type of data. Transcriptions of the speech of a group of twenty-five low-grade feeble-minded children in a state institution were attempted. Unfortunately a sufficient sample of sounds could be secured for only thirteen of these children. A considerable amount of time is necessary to make a transcription on each child and many of them produce sounds only after long intervals. Frequently no vocalization occurs. The work was done during the first week of the year 1941. A second transcription of the sounds of these children was obtained on December 22, 1941. Approximately, then, a year had elapsed between the taking of the two samples. During the year the number had been further reduced by death to ten cases. It is recognized that ten cases do not constitute an adequate sample for sound statistical treatment; nevertheless, there are tendencies in these data which possess at least preliminary meaning, and an elaborate statistical treatment will be postponed pending the accumulation of more cases. The method of sampling the speech sounds of immature children has been described elsewhere (4). Briefly, it consists of such sounds as occur on a given number of breaths of each subject. In the present study 30 breaths constituted the sample from each child. The vowel sample on the first transcription consisted of 417 elements. On the second transcription it amounted to 560. The total thus is close to 1,000 vowel sounds. On the first transcription 174 consonant elements were included; on the second, 332. The total consonant sounds used by these children thus amounted to about 500. Altogether, about 1,500 separate speech elements are available for analysis. Transcriptions were taken in the International Phonetic Alphabet as adapted by Fairbanks (1). The problem of the reliability of the observer and of the data in transcriptions of the speech of immature children has been discussed by Irwin and Curry (4), Irwin and Krehbiel (5), and Irwin and Chen (3). At the beginning of the year these children ranged in age from about