The measurement of waist circumference (WC) is widely advocated as a simple anthropometric marker of health risk; yet there remains no uniformly accepted protocol. This study determined whether the magnitude of WC differs across four measurement sites, and quantified the influence of site on the apparent prevalence of abdominal obesity. The predominantly white sample consisted of 223 men and 319 women (20-67 years). WC was measured using a nonstretching tape at the superior border of the iliac crest, midpoint between the iliac crest and lowest rib, umbilicus, and the minimal waist. Differences in WC across sites were tested using repeated measures ANOVA, adjusted for multiple comparisons. Inter- and intraobserver reliabilities across sites were estimated using intraclass correlation. In women, the mean WC for all sites were significantly different from each other, with the exception of the iliac crest and midpoint. In contrast, no significant differences between sites were found in men. Measurement site had an influence on the apparent prevalence of abdominal obesity (>88/102 cm), ranging from 23 to 34% in men and 31 to 55% in women. The reproducibility of WC was high at all sites and was comparable across levels of BMI. In conclusion, the magnitude of WC is influenced by measurement site, particularly in women. Small differences are amplified when dichotomous cut points rather than a continuum are used to define abdominal obesity. Adopting a standard measurement protocol will facilitate the interpretation and clinical utility of WC for obesity-related risk stratification.