Though it has been established that skinfold anthropometry has severe
limitations as a method of deriving total body fat (TBF), the possibil
ity 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 + (H2O)-H-3 dilu
tion), suggesting that the latter may perhaps be a better standard tha
n densitometry. (C) 1998 Elsevier Science Ltd. All rights reserved.