A model is presented which will predict mean age at first egg (AFE) for pullets of laying strains reared under non-limiting environmental conditions but exposed to a single change in photoperiod during the rearing stage. An initial analysis of 12 previously reported trials involving a wide range of genotypes showed that the response to an increase in photoperiod is not simply the inverse of the response to an equal decrease in photoperiod applied at the same age. Maximum sensitivity to a reduction in photoperiod was found shortly before onset of lay, whereas maximum sensitivity to an increment in photoperiod was observed at around 10 weeks of age. Two experiments were conducted to provide further data. The first compared the effect of 3-h increases in photoperiod from 8 h to 11 h or from 11 h to 14 h with the double increment from 8 h to 14 h and also tested a reduction from 11 h to 8 h, all imposed at 17 weeks of age. AFE was advanced to a similar extent by the changes from 8 to 11 h and from 11 to 14 h (9.8 and 10.9 days respectively). Response to the double increment was not additive: AFE on this treatment was 13.3 days earlier than for constant 8 h controls. Reduction in photoperiod from 11 to 8 h at 17 weeks delayed AFE by 18.7 days compared with constant 11-h controls. In the second experiment, pullets of two strains were transferred from 8 to 16-h photoperiods and from 16 to 8 h at 5, 7, 9, 15, 17 and 19 weeks of age. Controls were kept on constant 8 and constant 16-h days. Transfer from 8 to 16-h photoperiods at 5 weeks of age had no effect on AFE. At 7 weeks there was a bimodal response with some pullets subsequently showing advanced maturity and others not. Maximum stimulation of early maturity (31 days on average for the two genotypes) was obtained at 9 weeks of age and response to stimulation declined linearly with age thereafter. The delay in AFE resulting from a reduction in photoperiod (16 to 8 h) increased linearly between 0 and 15 weeks. At 17 and 19 weeks, the response was bimodal, with some pullets maturing at the same age as long-day controls and others showing delayed maturity. Using all this evidence and some other unpublished data, a model is developed to predict AFE as a function of mean photoperiod and change in photoperiod during the rearing phase. Elements are incorporated to allow for the insensitivity of pullets younger than 50 days to an increase in photoperiod and the effect observed late in rearing when a change in photoperiod comes too late to alter AFE for the most precocious individuals in a flock. Two coefficients are required to adjust for genotype. One describes mean AFE for the genotype when reared on constant daylength and the other defines the rate at which age effects the response to a single change in photoperiod.