The aim of the present study was to evaluate air-displacement plethysmography (ADP) in children aged 5-7 years. Body-composition measurements were obtained by ADP, (2)H dilution and anthropometry in twenty-eight children. Calculation of body volume by ADP was undertaken using adult and children's equations for predicting lung volume and surface area. Fat-free mass (FFM) was calculated using a three-component model. Measured FFM hydration was then compared with values from the reference child. Differences between measured and reference hydration were back-extrapolated, to calculate the error in ADP that would account for any disagreement. Propagation of error was used to distinguish the contributions of methodological precision and biological variability to total hydration variability. The use of children's equations influenced the results for lung volume but not surface area. The mean difference between measured and reference hydration was 0.6 (sd 1.7) % (P<0.10), equivalent to an error in body volume of 0.04 (sd 0.20) litres (P<0.30), and in percentage fat of 0.4 (sd 1.9) (P<0.28). The limits of agreement in individuals could be attributed to methodological precision and biological variability in hydration. It is concluded that accuracy of ADP was high for the whole group, with a mean bias of <0.5 % fat using the three-component model, and after taking into account biological variability in hydration, the limits of agreement were around +/-2 % fat in individuals. Paediatric rather than adult equations for lung volume estimation should be used.