This paper examines the familial aggregation of physician-diagnosed asthma in relation to the age- and sex-standardized total serum IgE levels of children and their parents in a sample of the general population in Tucson, Arizona, that has been followed in a longitudinal study for over 20 yr. There were 591 nuclear families containing 1,177 children who provided information about the presence or absence of a physician diagnosis of asthma. The serum IgE data were less complete: both parents and one or more of their children in 251 of the nuclear families, containing 468 children, had serum IgE levels measured. There was a very strong tendency for asthmatic patients to have asthmatic children, but only a small part of this appeared to be related to the familial aggregation of total serum IgE. In the absence of an asthmatic parent, there was a slight but significantly higher prevalence of asthma in children of whom both parents had IgE levels in the highest tertile. Very high rates of children's asthma depended on there being an asthmatic parent who also had at least moderate levels of serum IgE. It was also shown that asthmatic children have considerably higher total IgE levels than would be expected on the basis of their parents' IgE levels alone. The data appear compatible with several familial-aggregation hypotheses and a strong environmental influence determining which children are likely to develop asthma. We speculate that the inflammation in the airways of asthmatic patients itself tends to increase the serum IgE level, possibly secondary to mediators that it generates.