Three experiments examined the importance of the time and duration of exposure to a long day followed by a short day photoperiod signal in initiating reproductive activity in ewes. In Expt 1, ewes were maintained on short days (8.5 h light:15.5 h dark) from 21 December interrupted with either 105 long days (18 h light:6 h dark; LD) from 9 February or 35 LD from 9 February, 16 March or 20 April. Exposure to long days followed by short days advanced the onset of reproductive activity in comparison to control ewes maintained on simulated natural photoperiod. Exposure to long days for 105 days delayed the onset of reproductive activity (August 2 +/- 3 days; P < 0.05) compared with 35 days beginning on the same date (July 13 +/- 5 days). The interval from the end of the long day signal to the onset of reproductive activity was shorter (P < 0.001) however, after 105 LD than after 35 LD. In Expt 2, control ewes were moved from natural photoperiod to simulated natural photoperiod on 1 November and subsequently exposed to short days from 21 December. Four other groups were also exposed to this basic photoperiodic signal sequence but it was interrupted with either 70 LD from 16 November, or 35 LD from 16 November, 21 December or 20 April. More ewes (P < 0.05) initiated reproductive activity after exposure to 70 LD from 16 November and 35 LD from 21 December or 20 April compared with control ewes maintained on short days or ewes given 35 LD from 16 November. The interval from the end of long days to the onset of reproductive activity was less (P < 0.01) in ewes given 70 LD than in ewes given 35 LD. In Expt 3, ewes on natural photoperiod were given either 90 LD from 21 September, 35 LD from 21 September, 26 October, 30 November, 4 January or 8 February followed by short days. The majority of ewes that received long followed by short days after the winter solstice resumed reproductive activity. However, all photoperiod signals given between the autumn equinox and the winter solstice failed to initiate reproductive activity in ewes during the experiment. Thus we conclude that, in ewes, the reproductive neuroendocrine axis is insensitive to long days followed by short days between the autumn equinox and the winter solstice. The reproductive axis of ewes regains sensitivity to the inductive effects of long days followed by short days at a time close to the winter solstice. Between the winter and summer solstices, long days followed by short days maintain the anoestrous state and provide the cue for initiation of reproductive activity.