For all the skeptics in the recent BMI thread, here is some more information about obesity and mortality:
Simoson et al, Obesity 2007;15:994-1003
Studied about 40,000 people (both sexes) in Australia (a different ethnic stew than the USA), tried to correlate measures of obesity with overall mortality. Waist measurement and waist-to-hip ratios were the strongest predictors (better than BMI). For practical purposes, the waist measurement alone is just about as good as the ratio.
You measure waist circumference at the upper edge of your side pelvic bones (iliac crest) and keep the tape horizontal regardless of posture. 40 and 35 inches (men and women respectively) are clearly an important danger point, but it's a linear relationship. This is not generally the way you measure clothing, by the way. A good practice is to calculate the BMI and use the waist measurement to dampen errors resulting from high muscle mass, unusual body types, etc.
Waist-to-hip ratio for men should be less than 1 and for women less than 0.85 ideally. Above those levels risk rises steadily.
So, boys and girls, if you don't believe in the BMI get out the tape measure.