The number of circulating eosinophils is associated with the risk of asthma in population samples. Therefore, eosinophil levels may be an intermediate phenotype for asthma amenable to genetic analysis. We examined familial aggregation of the number of eosinophils x 10(6) L(-1) and the percentage of eosinophils based on a 300 count differential in 644 Hispanic and non-Hispanic white families, with 2, 097 subjects, enrolled in the Tucson Children's Respiratory Study. Both measures were adjusted for age, season and year at the time blood was drawn, sex, and ethnicity. Segregation analysis was conducted in the 458 non-Hispanic white families, as there were no significant familial correlations in the Hispanic families, and there was significant heterogeneity by ethnic group. Familial correlations (rho) in the non-Hispanic white families were as follows: mother-father, 0.05; mother-child, 0.18 (p < 0.001); father-child, 0.07; sibling-sibling, 0.31 (p < 0.001). Without covariates analyses indicated a polygenic/multifactorial mode of inheritance. After adjusting for current and past asthma an oligogenic mode of inheritance was suggested, plus additional residual familial components that were mainly maternally mediated. This study supports the notion of multiple, relatively common genes interacting to determine genetic susceptibility to asthma. Holberg CJ, Halonen M, Wright AL, Martinez FD. Familial aggregation and segregation analysis of eosinophil levels.