The symptomatology of infections as well as immunological and virological findings were analysed using a logistic model in a survey of 58 children experiencing their first febrile seizures. These were then compared with findings in 116 age- and sex-matched controls with infections but no seizures. There were no statistically significant differences in the aetiology of infections between patients and controls. High temperature was the only variable to explain the occurrence of febrile seizures in the logistic model after adjusting for duration of symptoms (partial correlation coefficient in logistic model, r = 0.31). The duration of symptoms before hospitalization was shorter in patients than in controls (mean 1.0 and 3.6 days). With a longer duration of symptoms, the likelihood of seizures diminished (r = -0.34). Patients in the seizure group had a significantly higher temperature at home than controls before hospitalization (39.4 versus 38.8 degrees C). Our findings of higher temperatures in children with febrile seizures supports its importance as the most important triggering factor in febrile convulsions.