A deterministic questionnaire method for zygosity determination is developed for use in epidemiological studies of adult twins. It is based on the answers of both members of a twin pair to two questions on similarity and confusion in childhood. The algorithm of the method is used to determine the zygosity status of a twin pair at two different levels of certainty. The validity of the method is tested by making blood marker determinations of 11 polymorphic marker systems fro a random sample of 104 twin pairs. The agreement between questionnaire and blood marker diagnosis was 100%, but the stricter level of certainty left 8.7% in the nonclassified group. The genetical representativeness of the sample is tested by the allele distribution of the markers as compared to the Finnish population data as well as by the distribution of the number of intra-pair differences in blood markers.