1. Turkey poults (1620) were used to compare the effects of three lighting programmes on heavy strain males reared to 188 d: constant light (24L:0D, CON); increasing light (6L:18D at 7 d gradually increasing to 20L:4D by 63 d, INC); a pattern identical to INC followed by a decrease in daylength from 84 d to 10L:14D at 112 d (DID). 2. Lighting affected growth pattern but had no effect on body weight at 118 d or overall food to gain ratio. 3. Both INC and DID lighting reduced overall mortality in comparison to CON light primarily because of a reduction in the incidence of skeletal disease and spontaneous cardiomyopathy. INC and DID lighting increased the incidence of cannibalism. 4. Turkeys given INC or DID lighting had a superior ability to walk in comparison to those birds given CON light. 5. INC and DID males stood, ate and drank more frequently, and sat less often than CON turkeys during behavioural observation. 6. There were no lighting effects on carcase composition except that INC and DID birds had heavier keel bones. The ultimate force per cm2 (stress) required to break femora was greater for turkeys given INC and DID lighting (P = 0.065). 7. Plasma testosterone concentrations at 117 d were 272.5, 115.2 and 29.5 pg/ml for turkeys given CON, INC and DID lighting, respectively (P = 0.072). Testosterone concentration was not related to growth rate.