White Leghorn strains were crossed reciprocally in a complete factorial mating producing 6 pure strains and 30 strain-crosses, which were kept for two laying cycles: 133–496 d of age and 547–909 d. Hens were housed for lay individually in four two-tiered batteries of cages. Strain additive effects (Ai), strain sex-linked effects (Zi), strain-cross heterotic effects (hij) and residual effects were calculated using regression. Viability was high in the first cycle of egg production with only 1 to 3% mortality in each of the four 11-wk periods, but lower in the second cycle decreasing with age. There was significant variation among strains in additive autosomal and sex-linked genetic effects and strain-cross heterotic effects, which increased with age in the second cycle. Heterosis for viability was positive in some strain-crosses and negative in others with considerable changes with age. The magnitude of heterotic effects was generally greater than the magnitude of additive or sex-linked genetic effects for viability. These results imply that different genotypes mount subtly different genetic responses to the problems of viability with advancing age and that more than one theory of ageing could apply. The results are discussed in relation to the theoretical aspects of ageing genetics. Key words: Ageing, fitness, viability, genetic effects, genetic variation, environmental variation