Manufacturer-supplied lumbar spine phantoms are normally used for quality control of dual X-ray absorptiometry (DXA) instruments. Presently, there is no pediatric phantom for whole-body mineralization and softtissue composition DXA measurements. We designed blocks of acrylic (for fat mass), polyvinyl chloride (for lean mass), and aluminum (for bone mass) whose combination provides five whole-body phantoms ("Inphants") that mimic body weight and composition during the first year of life and help solve problems that require repeated scans in stable conditions. Inphants were scanned using an Hologic QDR 2000. Comparisons were made between values obtained with and without the table pad, using infant software. Then we compared data obtained using infant and adult softwares successively in the same phantoms. The table pad significantly influenced DXA measurements. We observed significant differences in fat mass (p = 0.04) and lean mass (p = 0.03) with the smaller Inphant (3 kg) and in bone mineral content (BMC) (p = 0.02) with the larger Inphant (13 kg). BMC was three to five times lower with adult than with infant software. Adult software yielded systematically significantly lower fat masses but higher lean masses than infant software. Because there was no overlap with larger Inphants, we calculated conversion formulae between values of infant and adult software. The results suggest guidelines for scan acquisition and analysis in young subjects.