There is a growing body of literature that describes the use of dual-energy x-ray absorptiometry (DXA) for bone mineral content (BMC) and fat mass (FM) assessment in neonates, but the reproducibility and accuracy of the method are still controversial. Two different software programs have been developed for use on Hologic densitometers: the Pediatric Whole Body (PWB) and the Infant Whole Body (IWB) programs. They differ in scan time, radiation exposure, and in the algorithm used to assess BMC. We evaluated the reproducibility and accuracy of PWB and IWB in newborn piglets. Reproducibility of body mass (BM), FM, and BMC measurements from PWB and IWB were similar. BM agreed well with scale weight with both software programs; IWB was within +/- 0.5% and PWB was within +/- 0.3% of scale weight. FM was highly correlated with carcass fat (PWB: r = 0.962; IWB: r = 0.980). Errors in the DXA estimation of fat were similar with PWB and IWB. With both software programs, BMC was highly correlated with carcass calcium (PWB: r = 0.925, IWB: r = 0.987), but errors in the DXA estimation of calcium were about twice as high with PWB (+/- 16.9%) than with IWB (+/-9.2%). In four piglets, the addition of a layer of porcine lard was associated with an increase in BMC; this effect was more pronounced with PWB (+ 156%) than with IWB (+ 15%). The IWB software provided BMC measurements that were more precise, accurate, and stable in the presence of added fat than the measurements obtained with PWB, indicating that IWB is superior to PWB for in vivo determination of BMC and body composition.