Total body water was estimated as part of the assessment of body composition in children with growth disorders, using the newly commercially available method of bioelectrical impedance. This was undertaken to compare the precision and accuracy of the results with those derived from skinfold thickness against measurement of stable isotopically labelled water (H2(18)O) dilution as a standard. The comparisons were carried out to see to what extent the impedance method could be applied with confidence to assessment of children with growth disorders. Total body water was derived from impedance (I) using an association with height (Ht2/I). Impedance and skinfold thickness estimates of total body water were equally precise when compared with values obtained from H2(18)O dilution (limits of agreement -1.9 to +1.3 and -1.7 to +2.0 kg respectively). The mean intraobserver coefficient of variation for repeat measurements of impedance was 0.9% compared with 4.6% for skinfold thickness with an interobserver coefficient of variation for impedance of 2.8%. Bioelectrical impedance estimation of body composition is likely to be of value in the growth clinic when expertise in measurement of skinfold thickness is limited or repeated measurements are to be undertaken by different observers.