Objective-To examine whether the observed excess of childhood leukaemia and lymphoma near the Sellafield nuclear plant is associated with established risk factors or with factors related to the plant.Design-A case-control study.

schoolchildren and diagnosed before 1971 which was the year when the National Cancer Registry first passed on details to the NHS Central Register. Since then three other cases of cancer have been reported in the follow up ofthese schoolchildren, which is about one more than the expected number (2104) of non-fatal cases, plus two cases of carcinoma in situ of the cervix uteri compared with 1P79 expected.Overall 10 deaths were reported among children in the schools cohort compared with 12-69 expected at national rates-a deficit of about 20%. Apart from the one death from cancer the remaining nine deaths were all certified to "injury and poisoning," which is the largest cause ofdeath nationally for the ages that these children have achieved.Of some concern in this study is the 93% trace rate of the schoolchildren in the NHS Central Register. The 7% shortfall may have occurred, for instance, because of inaccuracies in the children's details that were recorded in the three school registers and difficulties in tracing entries in the Central Register from deficiencies in its construction during the early birth years of this study. It is not known whether these deficiencies, which resulted from parents not applying for welfare foods, come from any specific subgroup of families. For this and other reasons we are continuing to investigate ways of successfully identifying the untraced follow up records.We thank the Registration Division of the Office of Population Censuses and Surveys, the former headmistresses of Singing Surf and Calder Girls Schools, the Cumbrian County Council Education Department, and the headmaster of Seascale School for their cooperation in identifying records on children included in the study. We also thank the National Foliow up study of children born to mothers resident in Seascale, West Cumbria (birth cohort) M J GARDNER, A J HALL, S DOWNES, J D TERRELL Abstract Records on 1068 children who were born to mothers resident in Seascale Civil Parish during 1950-83 were studied. There was a large degree ofmobility among the families, and nearly halfofthe children did not subsequently attend the main local school. Use of the National Health Service Central Register, however, enabled us to follow up the children's records regardless ofplace of residence. The excess of leukaemia among Seascale children first supported from the analysis of geographical areas is confirmed. There were five deaths from leukaemia identified to 30 June 1986 compared with 0*53 expected at national rates-a ratio of9-36 (95% confidence interval 3*04 to 21.84). One ofthese deaths occurred after the child had left Seascale. There were four deaths from other cancers compared with 1-06 expected-a ratio of 3-76 (95% CI 1.02 to 9.63). In addition, three further cases of cancer, apart from the deaths, were reported compared with 1-19 expected since 1971-a ratio of 2-53 (95% CI 0*52 to 7.40).For other causes of death, including stillbirths and infant

Records on 1546 children who were identified as-having attended schools in Seascale up to November 1984 and were born since 1950 but not in the civil parish were studied. These children lived in or near Seascale for a period oftime while they were attending one or more of three local schools and are an additional group to the 1068 children who were identified as born to mothers resident in Seascale in an accompanying study. Even though some of the schoolchildren apparently remained in the village for a short period only all but 7% were followed up through the National Health Service Central Register. Mortality among these children to 30 June 1986 is comparable to that expected at national rates. From all causes there were 10 observed deaths compared with 12-69 expected-a ratio of 0-79 (95% confidence interval 0-38 to 1-45)-and from cancer one observed death compared with 2.04 expected-a ratio of 0-49 (95% CI 0-01 to 2.73). No deaths from leukaemia or lymphoma were reported, but only 0-83 was expected. Since 1971 (the year when cases of cancer were first notified to the NHS Central Register) three non-fatal cases of cancer were reported, including two lymphomas, compared with 2-04 expected and two cases of carcinoma in situ of the cervix compared with 1-79 expected. In addition, there was a case of