In two headache questionnaire surveys we inquired about the occurrence of headache in the mothers, fathers, siblings and children of the respondents. In total, 633 people completed valid questionnaires, 260 in the first survey and 373 in the second. The hypothesis was that familial headache occurrence would be positively associated with headache frequency. In each survey, the regression of headache frequency on the number of parents having headache was highly significant. Neither sex nor the sibling and children variables were significant predictors. In the cross-tabulations of the parental occurrence of headache with headache frequency we saw a clear "break-point" between the "no headache" and the headache frequency categories studied. For the final analyses the dichotomy "headache/no headache" was related in fourfold tables to headache occurrence in the father and the mother separately, and to the number of headache parents. The positive associations were not simply due to the large number of migraine cases since they remained after removing the migraineurs.