This study was performed to determine the precision and stability of dual-energy X-ray absorptiometry (DEXA) measurements, to compare bone mineral density (BMD) of subjects measured by DEXA and radionuclide dual-photon absorptiometry (DPA), and to evaluate different absorber materials for use with an external standard. Short-term precision (% coefficient of variation, CV) was determined in 6 subjects scanned six times each with repositioning, initially and 9 months later. Mean CV was 1.04% for spine and 2.13% for femoral neck BMD; for whole-body measurements in 5 subjects, mean CV was 0.64% for BMD, 2.2% for fat, and 1.05% for lean body mass. Precision of aluminum phantom measurements made over a 9-month period was 0.89% with the phantom in 15.2 cm, 0.88% in 20.3 cm, and 1.42% in 27.9 cm of water. In 51 subjects, BMD by DEXA and DPA was correlated for the spine (r = 0.98, P = 0.000) and femoral neck (r = 0.91, P = 0.000). Spine BMD was 4.5% lower and femoral neck BMD 3.1% higher by DEXA than by DPA. An aluminum phantom was scanned repeatedly, in both water and in an oil/water (30:70) mixture at thickness ranging from 15.2 through 27.9 cm. Phantom BMD was lower at 15.2 cm than at higher thicknesses of both water and oil/water (P = 0.05, ANOVA). The phantom was scanned repeatedly in 15.2, 20.3, and 27.9 cm of water over a 9 month period. In 15.2 and 20.3 cm of water, phantom BMD did not vary significantly whereas in 27.9 cm of water (equivalent to a human over 30 cm thick), phantom BMD increased 2.3% (P = 0.01) over the 9 months.