Only in the last few years has the scientific study of hormonal replacement therapy for hyposexuality begun in earnest with the advent of appropriately controlled experiment studies. Dose-response relationships can be demonstrated between testosterone (T) and sexual measures, but these have not yet been investigated in detail. Some aspects of sexual function are maintained in the presence of androgen levels well below the normal range, but preliminary evidence suggests that within a normal population high levels of T are correlated with more vigorous responses to visual erotic stimuli. Though T (and to a greater extent free T) declines with aging in parallel with the decline of sexual function, these hormonal changes contribute only to a minor extent to the behavioural change. Some non-aromatizable androgens may be less effective in stimulating sexual behaviour than T, but initial data on effects of dihydrotestosterone suggests that the capacity of an androgen to be aromatized (converted to oestrogen) is not a requirement for its sexual action. While T apparently increases the incidence of all types of male sexual activity, recent data contradict the belief that it directly facilitates the erectile mechanism in men, even though erection frequency is greatly reduced in untreated hypogonadal men. At the present juncture, it appears that the initial action of T may be on libido factors which lead in turn to the stimulation of other aspects of sexuality. Specifically, we propose that androgen acts through stimulating genital sensations and/or other pleasurable awareness of sexual response rather than directly through cognitive processes such as sexual imagery.