The most widely used measure of adiposity is to express weight adjusted for height using the body mass index (BMI). However, its limitations such as its inability to distinguish muscle weight from fat weight are well known, leading public health authorities in the UK and US to recommend measuring waist circumference as a complementary diagnostic tool for obesity. Recent attention placed on the syndrome referred to as 'normal weight obesity' – individuals with normal BMI but high body fat content – emphasizes the need for a more comprehensive diagnostic tool for obesity. Based on the NHANES III data, we utilize a semi-parametric spline approach to depict graphically the relationship between BMI, waist circumference and percent body fat. In this note, we propose that percent body fat charts that incorporate information from three anthropometric dimensions supersede the one-size-fits-all obesity diagnostic approach based on power-type indices such as the BMI.