It was previously found that dual-energy X-ray absorptiometry (DEXA) underestimated central body fat. The purposes of this study were to determine whether an updated version (enhanced version 5.64) of the analysis program corrected this problem (experiment 1) and to compare body composition assessed by DEXA and hydrodensitometry (HD) in women (n = 225) and men (n = 110) across a 21- to 81-yr age range (experiment 2). For experiment 1, 10 subjects underwent DEXA procedures in a control condition and with packets of lard positioned over either the thighs or the truncal region. DEXA accurately quantified the additional mass as approximately 96% fat, regardless of position. For experiment 2, DEXA yielded higher (P < 0.001) estimates of fatness than did HD (32.1 +/- 12.0 vs. 31.2 +/- 10.1%). The mean difference between the two methods was similar in young, middle-aged, and older subjects, but was different in men (HD-DEXA, 1.6 +/- 3.4% of body wt) than in women (-2.1 +/- 3.8% of body wt). Correcting the density of fat-free mass for variance in the bone mineral fraction of fat-free mass reduced the difference between the methods in men from 1.6 +/- 3.4 to -0.7 +/- 2.9% but widened it in women from -2.1 +/- 3.8 to -3.5 +/- 3.4%. A second correction procedure that adjusted for variance in water, protein, and mineral fractions of fat-free mass eliminated the differences in estimates of fat content by DEXA and HD in both men (21.1 +/- 9.3 vs. 20.6 +/- 8.4%, respectively) and women (37.5 +/- 9.3 vs. 36.8 +/- 8.0%, respectively). These results provide encouraging, but not definitive, evidence that the assessment of body composition by DEXA is accurate under the specified conditions.