The established Cole-Cole and newer Siconolfi methods for analysis of multifrequency impedance data were compared in a group of normal healthy individuals studied under bedside conditions. Impedance quotients derived from each procedure were similarly well correlated with independent dilutional estimates of total body water. Although both methods estimated resistance at zero frequency, were highly correlated, and provided impedance quotients with very similar correlations with extracellular water, these estimates were significantly different and thus may not be used interchangeably. The Siconolfi procedure demonstrated no significant advantage over the Cole-Cole method. In view of the sound theoretical basis for the latter, it is concluded that analysis of MFBIA data by the Cole-Cole method is to be preferred.