The relationship of stature with the prevalence of 18 chronic diseases or groups of diseases was analysed using data from the 1983 Italian National Health Survey, based on a sample of 63,859 individuals aged 20 or over randomly selected within strata of geographical area, size of the place of residence and of the household in order to be representative of the Italian population. Rate ratios (RR) were computed using multiple logistic regression, including terms for sex, age, geographical area, education and smoking. For 15 out of 18 diseases or groups of diseases the RR was below unity in the highest quartiles of height, and the inverse trends with stature were significant for 11 (diabetes, RR 0.90 for highest vs lowest quartile; heart disease, RR 0.92; chronic bronchitis and emphysema, RR 0.84; bronchial asthma, RR 0.70; anaemias, RR 0.70; liver cirrhosis, RR 0.62; urolithiasis, RR 0.76; renal insufficiency, RR 0.71; arthritis, RR 0.89; psychiatric and neurological disorders, RR 0.82). None of the diseases considered showed significant direct trends with height, but hypertension (RR 1.09 for the highest vs lowest quartile), haemorrhoids or varices (RR 1.09) and cancers (RR 1.22) tended to be elevated in the highest quartile of height. The generalised inverse relationship between height and prevalence of chronic disease suggests that poorer nutrition in childhood and adolescence is an unfavourable indicator for the subsequent occurrence of several diseases. Major exceptions were hypertension and varices, two conditions highly dependent on the pattern of health care utilization, and cancer.