Twelve subjects with apraxia of speech and minimal aphasic involvement were tested in four experimental conditions: effects of instructions, the effect of three different experimentally imposed response-delay intervals on a word-repetition task, the effect of noise, and the effect of visual monitoring. Also studied in one or more of these conditions were the loci of errors in oral reading, the apraxic subjects' ability to predict and to recognize their errors, and the nature of the errors made. Under the conditions in which they were studied, instructions, response-delay intervals, noise, and visual monitoring had no significant influence on phonemic accuracy. Subjects with apraxia of speech had significantly more difficulty with three- syllable words than with one-syllable words. They made more errors on words weighted high (Brown’s word-weighting method) than on words weighted low; word length and grammatical class appeared to be important characteristics influencing increases in errors. The ability of apraxic subjects to predict errors appears to be an individual characteristic; the ability to recognize errors appears to be a group characteristic. Subjects consistently made substitution, repetition, addition, and omission errors. The results support the contention that apraxia of speech is a motor programming disorder.