To our knowledge, this is the first population-based case-control study of risk factors for absence seizures (AS). Diagnosis of AS was based on clinical criteria. The complete medical history of potential cases, available through the records-linkage system for residents of Rochester, MN, was independently reviewed by three neurologists who agreed upon the diagnosis. All AS patients who were residents of Rochester at time of diagnosis between 1935 and 1979, and who were born in this community, were included (N = 30). Two population controls (born in Rochester) were matched to each patient, and for both patients and controls, the records-linkage system was used to obtain information about possible risk factors. The only factor significantly more common in cases than in controls was a history of febrile seizures (odds ratio = 12; p less than 0.01). We suggest that these febrile seizures represent either an early manifestation of the convulsive diathesis or the symptom of a preexisting brain dysfunction. None of the other factors investigated reached statistical significance, including those that have been previously suggested such as twin pregnancy, breech presentation at delivery, being first-born, and perinatal asphyxia. Sample size limitations should be considered in interpreting these findings.