EDITOR, - In 1990 Gardner et al identified an association between paternal preconceptional irradiation and childhood leukaemia and nonHodgkin's lymphoma in a sub population of workers at sellafield living in west Cumbria.1 The Gardner hypothesis asserted that this association might be casual. In her editorial Hazel Inskip cites a series of studies that have been conducted to test the hypothesis.2 Perhaps the most telling of these is the recently published work at Parker et al, which tested the Gardner theory in the whole of the population from which it was derived.3 This study, along with the others described by inskip, have failed to support the hypothesis of paternal preconceptional irradiation, as inskip acknowledges.

Also mentioned is the possibility that paternal preconceptional irradiation might be a marker for some other occupational exposure. Although these other occupational factors should more properly be seen as confounders of the original study, some people have raised them to the status of surrogate Gardner hypothesis. Now these two have been tested by the Health and Safety executive's recent study and found not to support a Gardner-type theory.4 We work as chief medical officers in the nuclear industry, and our first duty is to keep employees fully informed of the risks …