A basic assumption of the doubly labeled water (DLW) and triply labeled water (TLW) methods for measuring water flux (rH2O), CO2 production (rCO2), and fractionated water loss (X) is that the H of body water only leaves the body as water. Any loss of isotopes in other products will introduce an error into these techniques. The body fat represents the largest potential sink for water H. 2H sequestration into the carcass fatty acids was investigated in eight pigs labeled with 2H2O for 21 days. rCO2 was measured simultaneously in respiration chambers to allow an accurate assessment of the effect of 2H sequestration on the estimated rCO2. The fat content of the diet (1.63%), level of intake, and stage of maturity were all designed to give the widest possible range of sequestration effects. Four animals were restricted to their estimated maintenance requirement and four were allowed to feed ad libitum giving a range of weight gain from 100 to 650 g/day. This was reflected in the estimated error on rH2O (+0.42% in the restricted group and +2.52% in the fast-growing animals) and on rCO2 (-1.30 and -7.59%, respectively). The error on the calculation of X using TLW was +0.03 units in the restricted group and +0.20 units in the fast-growing animals. The error of +0.2 on X propagates through to an underestimate of rCO2 of approximately 4%, and since this is additive with the error on DLW the ultimate error on rCO2 using TLW would be approximately -12%.(ABSTRACT TRUNCATED AT 250 WORDS)