The present study validates the use of dual energy X-ray absorptiometry (DEXA) for measurement of body composition. The precision error was expressed as the SD (CV%) for fat mass, FAT%, lean tissue mass, and total body bone mineral: 1.1 kg (6.4%), 1.6% (5.7%), 1.4 kg (3.1%), and 0.03 kg (1.2%), respectively. The accuracy study in vitro used (1) mixtures of water and alcohol, (2) mixtures of ox muscle and lard, and (3) dried bones. In the clinically relevant range of values there were only small influences on DEXA measurements of variations in amount and composition of the soft tissue equivalents. The accuracy study in vivo compared the components of body composition measured recently by DEXA and earlier by dual photon absorptiometry, counting of naturally occurring total body 40K, and body density by underwater weighing in 25 healthy adult subjects. We found agreement between fat percentage (and lean body mass) by DEXA and the three established measurements modalities; mean differences were (-5.3 to -0.4%) and (-0.7 to 2.5 kg) for fat percentage and lean body mass, respectively. We conclude that DEXA provides a new method of measuring body composition with precision and accuracy errors, which are compatible with the application of DEXA in group research studies and probably also in clinical measurements of the single subject.