This study was aimed at verifying whether accuracy and reproducibility of mineral content measurements of bone phantoms by single-photon absorptiometry change as a function of source strength, and whether these changes depend on the phantom's width. To these purposes, the Norland 2780 bone mineralometer was used. With this unit, the mineral contents of two cylindrical phantoms of different diameter and of known mineral content were measured every week, 5 times on the whole, up to exhaustion of the source. Both accuracy and reproducibility decreased as source strength decreased. The lowest reproducibility was observed for the measurement of the smaller phantom with low source strength. It was therefore concluded that when source strength reaches the threshold value of 2000 cps, the measurement is so much affected as to jeopardize the reliability of short-term longitudinal studies of bone mineral content in patients.