Plasma testosterone levels in a group of 341 men with sexual dysfunction were compared to those in 199 men with normal sexual function. All subjects were participants in a 2-week intensive conjoint sex therapy program at the Masters & Johnson Institute. Testosterone determinations were made using radioimmunoassay methods after column chromatography; all blood samples were obtained on the second day of therapy between 8:00 and 9:00 a.m. after an overnight fast. Circulating levels of testosterone in men with normal sexual function (mean 635 ng/dl) were not significantly different from testosterone values in sexually dysfunctional men (mean 629 ng/dl). However, men with primary impotence (N = 13) had significantly higher testosterone levels than men with secondary impotence (N = 180), with mean levels of 710 and 574 ng/dl, respectively (p < 0.001). The mean testosterone level for men with ejaculatory imcompetence was 660 ng/dl (N = 15), while for men with premature ejaculation the mean was 622 ng/dl (N = 91). Plasma testosterone concentrations were not related to therapy outcome but were correlated negatively with age of patients.