Though it has been established that skinfold anthropometry has severe limitations as a method of deriving total body fat (TBF), the possibility that the problem might be related more to the assumptions implicit in densitometry has to be addressed. This paper suggests that smaller residual errors might be obtained if sums of skinfolds are regressed against TBF measured by a difference technique (IVNAA + 3H2O dilution), suggesting that the latter may perhaps be a better standard than densitometry.