A case control study of 97 persons under 25 diagnosed with leukaemia or lymphoma from 1950-85 resident in West Cumbria, compared with 1001 age and sex matched controls has been carried out. Raised odds ratios have been found for being born; within five kilometres of the Sellafield nuclear plant; to mothers aged greater than 40 years at the time of birth and to fathers accumulating greater than 100 milliSieverts of radiation before conception. Raised odds ratios were also found for pre-natal radiography and viral infection, but the prevalence of these factors was not increased in the vicinity of Sellafield. Examination of habits of children that may have resulted in enhanced exposure to the Sellafield discharges showed no significant raised odds ratios. However, for cases of leukaemia eating shellfish more than weekly compared with less than monthly, an odds ratio of 6.00 (95% Confidence interval 0.50-71.7) was found based on two exposed cases, neither of whom were born within five kilometres of Sellafield. The raised odds ratio of 6.24 (95% Confidence Interval 1.51-25.76) found for preconceptual radiation dose in excess of 100 milliSieverts compared with zero dose, was based on four cases of leukaemia, three of whom were born within five kilometres of the Sellafield plant. For the other two cases of leukaemia born within 5 km of the plant one father had a radiation dose of 97 milliSieverts whilst the other father was known to have been employed at Sellafield at the time of birth of the index case, but his radiation dosimetry was not available. Therefore the risk of leukaemia for being born within five kilometres of Sellafield can be explained statistically by the association with paternal preconceptual radiation exposure of fathers of cases. This result has important implications for radiobiology and radiation protection.