A weight-height index of adiposity should indicate the relative fatness of subjects of differing height unless obesity is itself correlated with height. The average body fat among adult women attending a hospital outpatient clinic for obesity was 40.5 percent of body weight. The height of an unselected series of 286 of these outpatients was found to be similar to that of the general population of women of similar age, which indicates that obesity in adult women is not significantly related to height. Body composition was measured by body density, body water and body potassium in a series of 104 female and 24 male subjects aged 14-60 years. In both sexes density, water and potassium gave progressively higher estimates of body fat (kg), and there was a significant difference between the values by different methods. The average of the estimates by these three methods was taken to be the 'true' value for each individual (F kg). Regression of F/H2 on W/H2 (Quetelet's index) gave a correlation coefficient of 0.955 for women and 0.943 for men. The deviation of the body fat estimated from Quetelet's formula from the 'true' value was not much greater than that when density, water or potassium were used as a basis for estimating body fat. It is concluded that Quetelet's formula is both a convenient and reliable indicator of obesity.