The possible infective origin of childhood leukaemia has been the subject of much recent attention.1 Fearet al 2 have taken up my suggestion that my necessarily limited examination of childhood leukaemia relative to levels of paternal occupational contact in the general population3 be extended and (as in my study) found no evidence of a general increase in risk when fathers have jobs involving contact with many people. However, in reporting my hypothesis they did not bring out that the principal question considered in my study did not concern the general population, but rather whether there was a relation with paternal occupational contacts “in population-mixing situations associated with an excess of childhood leukaemia”.3 I investigated this question3with data from previous …