BACKGROUND: Recent epidemiological studies consistently report an inverse association between sibship size and allergic disease, but evidence from individuals born before the 1980s is inconsistent. As information on relative permanence of this finding may offer clues to its biological explanation, the association between sibship size and allergic disease in individuals born between 1918 and 1952 was investigated. METHODS: Cross sectional surveys conducted by the Student Health Service at the University of Glasgow (1948-68) provided data on 14 140 men and women aged 16-30 years at the time of examination. The main outcome measures studied were self-reported asthma, eczema-urticaria, and hay fever. RESULTS: A total of 1677 individuals (11.9%) provided a positive history of at least one of the three allergic diseases: 457 (3.2%) asthma, 594 (4.2%) eczema-urticaria, and 885 (6.3%) hay fever. Compared with those without siblings (reference odds ratio = 1), the odds ratios (95% confidence intervals) for having any allergic disease among those with one, two or three siblings were 0.86 (0.75 to 0.99), 0.80 (0.69 to 0.93), and 0.70 (0.60 to 0.83), respectively (p(trend)<0.001). Increasing birth order and low socioeconomic position in childhood were associated with a lower risk of allergy. Adjustment for birth order, year of birth, age, sex, socioeconomic position in childhood, and family history of allergy did not materially alter the results. CONCLUSIONS: There is a robust inverse association between sibship size and allergic disease even among people born in the first half of the 20th century. These results favour relatively time-independent explanations for this phenomenon (such as the hygiene hypothesis or parity related changes in the intrauterine environment) over new environmental exposures.