We have developed a method for measuring true fractional calcium absorption (α) in premature infants using two stable isotopes of calcium and tested it in seven studies in seven infants (birth weight 1543±65 g, gestation 32.8±7 wk). A total of 7.5 μg/kg 46 Ca was given as a single intravenous bolus. Immediately thereafter 1.25 mg/kg of 44 CA was given in a single gavage feeding of standard infant formula (Enfamil). A metabolic isolette was used to obtain 4-h collections of urine for 24 h total. 46 Ca and 44 Ca were measured in urine by thermal ionization mass spectroscopy and expressed as the ratio to naturally occurring 48 Ca. The differences in the 46 Ca/48 Ca and 44 Ca/48 Ca ratios from natural levels (Δ% excess 46 Ca and Δ% excess 44 Ca) were calculated. Percent absorption (α) equals a constant times cumulative Δ% excess 44 Ca/A% excess 46 Ca. The calculation of α is independent of urine volume or concentration. The Δ% excess 46 Ca, showed the expected multiexponential decline as a function of time, and Δ% excess 44 Ca usually peaked during a 4− to 8-h urine collection. Calculations of α using increasingly long sampling times showed that a plateau had been reached by 12 h. α values calculated after 16–24 h in the seven infants at 2 wk of age were 41, 48, 45, 46, 25, 55, and 51%. Repeat studies at 3 wk of age were 46, 60, and 54%. These values are somewhat higher than net percent calcium absorption values reported for standard formula and thus appear very appropriate. This methodology will be very valuable in studying factors that may affect true calcium absorption in premature infants.