This study hypothesizes that certain types of laryngeal lesion can be identified by spectral analysis and computer evaluation of recorded speech samples. An experiment, using an acoustic-diagnostic technique and a computer is developed for identifying some of the more serious laryngeal anomalies. A pitch perturbation factor (PF) is derived from the experiment for each subject, and computer generated graphs are made to show the correlation between PF's and diagnostic categories of laryngeal lesion. Significant relationships are found between selected PF's and the following pathological conditions. (1) Carcinoma of the vocal cords. (2) Polyps on or near the vocal cords. (3) Acute and chronic laryngitis. © 1969.