A SIGNIFICANT contribution is achieved by Drs. Armitage, Court Brown and Doll in evaluating some models postulated in assessing the leukaemogenic potency of radiations. Along this line, it appears relevant to discuss some possible interpretations of a linear dose-response, observed over a limited range of leukaemia incidence, as evidenced in the study of patients irradiated for ankylosing spondilytis. Basic assumptions of some theoretical models will be first examined.