One of the problems in the design of continuous speech recognition systems is evaluation of the stage which detects occurrences of phonemes in the speech signal. A method of evaluating and optimizing phoneme recognition networks has been developed which uses data obtained by manual “marking” of phoneme locations on a tape recording of speech. With the aid of a digital computer and display, 1090 occurrences of vowels, nasals, and “don't care” conditions were marked in a five‐minute sample of speech. These data, which may be used for measurement of error rates in speech recognition systems, have been included in the IEEE Computer Society repository for data bases.