OBJECTIVES: (1) To compare 2-, 3-and 4-component models of body composition based on total body water (TBW), underwater weighing (UWW), skinfold thicknesses (SF), total body potassium (TBK), dual-energy X-ray absorptiometry (DXA) and total body electrical conductivity (TOBEC); (2) to compare postpartum changes in body composition estimated by the 2-, 3-and 4-component models and (3) to test for an effect of pregnancy or lactation on the hydration, density and potassium content of fat free mass (FFM) in postpartum women. DESIGN: Longitudinal measurements of body composition at 3, 6 and 12 months postpartum. SUBJECTS: Thirty-®ve healthy postpartum women, aged 30.263.5 y. MEASUREMENTS: Body composition was estimated by 2-component models based on TBW, UWW, SF, TBK, DXA or TOBEC; 3-component models based on TBW and UWW (Fuller 3, Siri 3); and a 4-component model (Fuller 4) based on TBW, UWW and bone mineral content. RESULTS: Systematic differences were seen among the various body composition models, with the following ranking from lowest to highest estimate of fat mass (FM): TOBEC, TBW, Fuller 3, Siri 3, Fuller 4, UWW, SF, TBK, and DXA. Estimated changes in FFM and FM were not signi®cantly different among methods, except for the 3±6 months FFM and FM changes estimated from TBW, which differed from SF, DXA, and TOBEC. Pregnancy-induced changes in the hydration, density and potassium content of FFM were not evident by 3 months postpartum (0.7360.02, 1.09960.015 kg/l and 2.3160.10 g/kg, respectively). CONCLUSION: In spite of systematic differences among body composition models for the measurement of FFM and FM, changes in FFM and FM did not differ signi®cantly among the models. Since there was no apparent effect of pregnancy or lactation on the postpartum composition of FFM, 2-component models of body composition are acceptable for use in postpartum women beyond the puerperium.