Two experiments investigated the response of the pituitary-gonadal axis of pinealectomized male Syrian hamsters to programmed systemic administration of melatonin. In the first experiment, castrated male Syrian hamsters were housed in a short photoperiod (8L:16D) and maintained on subcutaneous testosterone implants for 7 weeks. These males were then pinealectomized or sham-pinealectomized and their testosterone capsules removed. Daily infusions of melatonin 250 ng/infusion) or its vehicle were administered for 3 weeks; infusion duration was long (11 or 12 hr) or short (6 hr). Measurement of serum luteinizing hormone (LH) following this 3-week period indicated that long-duration melatonin infusions mimicked short-day conditions (LH levels were low), but short-duration infusions did not (LH levels were significantly elevated). In the second experiment, pinealectomized, gonadally intact males were housed in a 12L:12D photoperiod and injected once daily with melatonin or its vehicle, either 3 or 5 hr after dark onset for 11 weeks. These times were chosen to coincide with the light:dark cycle phase that according to published reports is optimally responsive to exogenous melatonin for the induction of short-photoperiodic effects. Melatonin injections did not induce gonadal regression in pinealectomized hamsters. Melatonin and vehicle-treated males responded similarly; their testis widths and serum testosterone levels were not significantly different at the end of the experiment. These results support the hypothesis that the duration of melatonin secretion each night is an important variable in conveying photoperiodic information, but that the circadian phase during which melatonin is present is not.