Bone density measurements are frequently carried out using photon absorptiometry techniques. The analysis of data collected in this way depends upon scatter-free detector signals. In practice the detector and source collimation lead to high levels of scattered radiation being detected. An analysis of the problem using photon transport computer models has shown that the current systems are reasonably insensitive to the scatter contribution except where patient changes occur over a long series of measurements or where new results taken with DEXA are compared with those taken on older systems such as DPA. This is frequently the case since the major body of data related to serial bone density measurements has been achieved with DPA. Inclusion of scattered radiation in the bone density calculations can lead to a 0.5-3.5% reduction between system types. Patient size changes could lead to a reduction in bone density of 0.5% (femoral neck measurement) to 1% (lumbar spine measurement).