Seltzer, C. C. (Dept. Nutrition, Harvard Univ. School of Public Health, Boston, Mass. 02115), H. W. Stoudt, Jr., B. Bell and J. Mayer. Reliability of relative body weight as a criterion of obesity. Amer. J. Epid., 1970, 92: 339–350.—The competence of the relative body weight ratio as an instrument of assessing obesity was investigated in a series of 1, 761 healthy U. S. Army veterans, based on the association of relative body weight values and skinfold measurements. The analysis indicates that the diagnosis of obesity by relative body weight ratios is notably poor. Only a minority of the members of the most extreme relative body weight categories of 120 and 125 or more are found to be frankly obese by skinfold standards, triceps, subscapular, or combined triceps and subscapular. These categories comprise nonobese and obese in varying proportions depending on the value of the ratio and the nature of the standard weight used in the calculation of the ratio. The implications of these results are far-reaching. Unwarranted generalizations may result from the use of the oversimplified relative body weight ratio as an index of obesity.