The present work is part of an ongoing study on how external and internal factors control the hypothalamus-pituitary-gonadal axis in male great tits (Parus major). In southwest Sweden adult and juvenile male great tits remain absolutely photorefraclory until mid/late September. The site of refractoriness lies in the hypothalamus or some higher neural circuit. Results from the present study suggest that refractoriness is induced by long days affecting the hypolhalamic content and/or secretion of GnRH through a testosterone independent mechanism. Photorefractoriness is both induced and maintained via a testosterone independent mechanism. Short days break refractoriness. From October and onward the postcastration LH response, as well as the photoperiodic LH and testicular responses in intact birds, increased progressively with season, and also with the number of short days to which the birds had been exposed. Maximum seasonal responses were attained in January. Contrary to photorefractory birds, low LH levels in photosensitive great tits are caused by the negative feedback effects of testosterone. This together with the fact that the postcastration response in short day treated birds increased with the number of short days the birds had been exposed to suggest that the reproductive system of the great lit actually is switched on by short days, and that the function ol long days in spring is more to reinforce the photoperiodic drive in order to override the negative feedback effects of testosterone. Furthermore, as the transition to a maximum photoperiodic response is a gradual process that takes several months the situation can be looked upon as a gradual breaking of photorefractoriness, i.e. after breaking absolute refractoriness the great tits are still partly long day refractory. This gradual recovery was found in birds exposed to normally decreasing daylengths (free-living birds) and in those given a constant exposure to short days for a year. The daylength the birds are exposed to determines the lime of onset of gonadal regression and decreased LH secretion. The duration of the entire cycle, as well as the amplitude of the LH/testicular response, is however determined by some other factor. Results indicate that the photoperiodic threshold may change with season. In early and mid autumn there were no increases in testosterone levels during a photoperiodic cycle. In late autumn testosterone levels increased dramatically after about 1 week of exposure to long days, and liters immediately rose when great tits were exposed to long days during late winter. Free-living great tils showed a small but significant increase in gonadal size and circulating levels of LH between December and January. Photorefractory great tits kept on short days lor a year showed similar increases between December and February. Despite the absence of long days testes in these latter birds developed secondary spermatocytes, but no spermalids or sperms. It seems likely that this winter increase is the result of endogenous rhythms. The present results indicate that observed discrepancies between species might in some cases be due to the lact that an experiment has been performed only at one time ol the year.