This review describes the advantages and limitations of dual-energy absorptiometry (DXA), a technique that is widely used clinically to assess a patient's risk of osteoporosis and to monitor the effects of therapy. DXA is also increasingly used to measure body composition in terms of fat and fat-free mass. There are three commercial manufacturers of DXA instruments: Lunar, Hologic, and Norland. All systems generate X-rays at two different energies and make use of the differential attenuation of the X-ray beam at these two energies to calculate the bone mineral content and soft tissue composition in the scanned region. Most DXA instruments measure bone mineral in the clinically important sites of the spine, hip, and forearm. More specialized systems also perform whole-body scans and can be used to determine the bone and soft tissue composition of the whole body and subregions such as arms, legs, and trunk. The effective dose incurred during DXA scanning is very small, and, consequently, DXA is a simple and safe technique that can be used for children and the old and frail. Precision of all DXA measurements is excellent but varies with the region under investigation. Precision is best for young healthy subjects (coefficient of variation is about 1% for the spine and whole body bone measurements) but is less good for osteoporotic and obese subjects. The accuracy of DXA measurements, however, can be problematic. Marked systematic differences in bone and soft tissue values are found between the three commercial systems due to differences in calibration, bone edge detection, and other factors. In addition, differences in reference data provided by each manufacturer can lead to an individual appearing normal on one machine but at risk of osteoporosis on another. At present, DXA cannot be regarded as a "gold standard" for body composition. However, the continuing development of DXA and the introduction of new software is greatly improving the performance of this increasingly important technique.