The distribution of weight in the adult population aged 20-69 years was examined by comparison of estimates obtained from the 1985 Health Promotion Survey and the 1981 Canada Fitness Survey. The Health Promotion Survey obtained information on self-reported weight and height, and the Canada Fitness Survey utilised measured weight and height. The classification of respondents into weight categories followed the recommendations of the 1973 Fogarty Conference on Obesity. Values of the Quetelet index defined as W/H2, where W = kilograms and H = metres, were used to define four weight categories: underweight, acceptable weight, overweight, and obese. The comparisons of prevalence estimates of the various weight categories indicate that self-reported weight and height leads to a systematic weight misclassification bias. The implications of this bias for epidemiological studies are discussed and suggestions are offered to handle the bias.