The purpose of this paper is to examine whether exposure to antigens in the early months of life can increase or decrease subsequent sensitization. With this aim in mind, 304 subjects undergoing hyposensitization were randomly selected in an out-patient study: 207 of them had pollenosis (seasonal asthma, rhinitis, and positive skin tests to grass pollen) and 97, house dust respiratory allergy (perennial asthma and rhinitis with positive skin tests to house dust mites). The results of the study show a statistically significant association (P less than .005) between birth in grass pollen season and pollenosis. The region of birth (north or south Italy) appeared to be a "deterministic" effect modificator; the odds ratio due to the birth season was 5.099 in the northern and 0.997 in the southern Italian regions. The main effect of the region of birth was not significant, neither was sex, age, and the region of birth of the parents (main, polinomial and interaction effects). It is likely that there is a real relationship between the month of birth of atopic subjects and the prevalence of the pollenosis. The precise reason(s) of the different degrees of association between month of birth and pollenosis in subjects born in northern and southern Italy is not clear.