In most mammalian species, reduced androgen availability is associated with marked reductions in male sexuality; conversely, androgen replacement in castrated males restores sex behavior within a few weeks. Testosterone (T) pulse duration, amplitude, frequency, and inter-pulse interval may be as important as total amount of hormone in determining target tissue responsiveness. We remain ignorant of the number and duration of daily T pulses necessary and sufficient to sustain male mating behavior. An in-dwelling infusion system was employed to vary T-pulse frequencies and durations. Daily 4 h infusions of aqueous T (100 microg/0.064 ml) and twice daily 4 h pulses of T (each 50 microg/0.064 ml) were sufficient to maintain ejaculatory behavior of sexually experienced castrated hamsters for 11 weeks post-castration; castrated hamsters infused with vehicle ceased to display the ejaculatory pattern 3 weeks after gonadectomy. Circulating T concentrations of hormone-infused hamsters declined markedly 7 h after the termination of each infusion. These results establish that male sex behavior can be sustained with infusions of relatively low T concentrations for 4 h/day and suggests that the basal concentrations of T sustained by the gonad during inter-pulse intervals may not be necessary for maintenance of sex behavior. 4 h T infusions were sufficient to maintain penile and seminal vesicles weights, but not ventral prostate weights or flank gland dimensions; the threshold for maintaining male sex behavior is lower than that for some androgen-dependent peripheral structures. Development of effective androgen replacement regimens that sustain sex behavior in castrated animals may be useful in the design of androgen replacement therapy for hypogonadal men.