In two municipalities in Lower Saxony statistically valid clusters were observed, which attracted great public interest. Committees were set up to initiate a large variety of on-the-spot-investigations. Finally, it was decided to conduct a case control study throughout Lower Saxony to explore potential risk factors which might explain the observed clusters. A limited number of already established and currently discussed hypotheses was chosen for investigation. The study was based on patients registrated at the German Children's Cancer Registry (GCCR). For each child with leukemia diseased between 1988 and 1993 two population-based controls (local and state controls) and one tumour control were selected. Data were collected by means of a questionnaire self-administered by the parents and a telephone interview. In addition, measurements of electromagnetic fields and radon were performed and inspections of the child's birth record were made for the purpose of confirming a potential association between parenteral vitamin K prophylaxis and leukemia or tumours. This paper presents the concept and basic considerations of the study, its design and statistical evaluation. Response rates and a summary of results will be presented, too. The paper will serve as a reference for subsequent publications about more detailed analyses of specific potential risk factors. 425 parents of diseased children and 610 of non-diseased children were asked for participation. The rates of response were 82% for families with diseased and 71% for families with non-diseased children. In total, 781 parents participated in the study. The most important results are as follows: The Greaves' hypothesis (9, 10) was supported by the following results: In children diseased with leukemia, vaccinations were less frequent, virus-related infections occurred more rarely, these children were more frequently first-born children and more frequently breast-feed, and they possibly had fewer contacts with other children in infancy. Our data do not show a significant association between parenteral vitamin K prophylaxis and leukemia or tumours (17). Measurements of electromagnetic fields show a weak association between high-level exposure and an increased risk of developing leukemias. Children who were X-rayed more than four times or were born prematurely and had received intensive care treatment show a positive association with occurrence of leukemia, but the number of these patients was very small and these factors are not independent. The only potential job-related hazard was paternal exposure to plastic and resin fumes. The incidence of miscarriages and abortions was increased in mothers of children with leukemia. Our study did not confirm the hypotheses, that leukemia is associated with gestational age, with consumption of alcohol, nicotine, and medicaments during pregnancy. Nor was any association detected with exposure to wood preservatives and insecticides or with a high socio-economic status. The number of patients living in municipalities with increased incidence of leukemia was too small to show statistically valid results. However, it is noteworthy that some of the above-mentioned risk factors were observed also in these children. One of the purposes of a nation-wide case control study which is currently performed at the GCCR is to validate and complete these results.