At the Bone Mineral Laboratory of the University of Wisconsin a microcomputer-based digital read-out system was designed specifically for determining bone mineral content (BMC) and bone width (BW) in newborn infants with the following features: (1) high accuracy and precision; (2) high reproducibilityin vivo; (3) direct read-out of BMC and BW; (4) automatic data calibration; and (5) use of a low activity [125I] source (r= 0.99). Short-term precision (weekly, coefficient of variation 1.7%) and long-term precision (monthly, coefficient of variation 2.1%) for measuring BMC were determined by multiple determinations on a four-chambered bone phantom calibrated with the bone sections. Immediate reproducibility (without repositioning the arm) for the 4–6 scans performed for each determination of BMC and BW was good with a mean coefficient of variation of 3.9% for BMC and 3.6% for BW. In 84 infants, repositioning error was determined by repeating the measurement of BMC and BW after repositioning the arm. The correlation coefficients between measurements before and after repositioning the arm were 0.97 for BMC and 0.95 for BW. BMC correlated well with gestational age (r= 0.92), birth weight (r= 0.89) and bone width (r= 0.92). BW also correlated with gestational age (r= 0.84) and birth weight (r= 0.85). A multiple linear regression analysis of BMC versus BW, gestational age, and birth weight was done. The correlation coefficient between the predicted BMC from these variables and measured BMC was 0.95. Photon absorptiometry can be used with high accuracy, precision, and reproducibilityin vivoin newborn infants. BMC correlates with gestational age, birth weight, and bone width.