The evaluation of bone by ultrasound is a new, transportable and non-irradiating method. It measures attenuation of ultrasound (BUA), as well as velocity (SOS) of ultrasound in a given tissue. For the measurement of bone, the most frequently used bone is the calcaneus. We have used this technique (Achilles, Lunar) to evaluate the reproducibility of measurements in healthy volunteers and to determine the correlations between the results of ultrasound measurements and those of conventional densitometry (DXA) in a population of postmenopausal females who consulted for screening or diagnosis of osteoporosis. Although several publications have already dealt with the same questions, it is important to reexamine them locally before the technique is uncritically used even in unspecialized centers. The coefficients of variation corresponded to that reported by others: 1.8% for BUA and 0.3% for SOS. The correlations between the ultrasound measurements and the results of DXA at lumbar spine, femoral neck and calcaneus in 57 postmenopausal women aged +/- 59 (43-78) years and in part osteoporotic (DXA results below 2.5 SD-T-score) (Hologic QDR 2000) were all significant (p < 0.001). The best correlation between the two techniques (DEXA and ultrasound) is found when applied to the same site, i.e. the calcaneus (r > 0.8). These results also agree with the literature. Low ultrasound values (BUA < 100 dB/MHz, SOS < 1495 m/s) are highly suggestive of osteoporosis, while high ultrasound values are almost exclusively found in non-osteoporotic women. This information provides the basis for further investigations, such as correlation with fracture risk, in cross-sectional and longitudinal studies.