Magnetic resonance imaging (MRI) has shown promise in the assessment of bone architecture. The precision and feasibility of MRI measurements in osteoporosis in vivo have been assessed in this study. T2' was calculated from measurements of T2 and T2* in the calcaneus of 32 postmenopausal women using a gradient-echo sequence PRIME (Partially Refocused Interleaved Multiple Echo). This sequence allows the measurement of T2 and T2* in one acquisition. In vivo measurements of bone mineral density (BMD) by dual-energy X-ray absorptiometry (DXA) were made in the calcaneus, spine and femoral neck. The ultrasound parameters broadband ultrasound attenuation (BUA) and speed of sound (SOS) were also measured in the calcaneus. These three techniques have not previously been compared in the same study population. The precision of the MRI technique was poor relative to the DXA and ultrasound techniques, with a CV of 6.9% +/- 4.4% for T2' and 5.5% +/- 3.6% for T2*. Approximately 4% of this is due to system error as determined by phantom measurements. The postmenopausal women were classified as having low BMD if they had a lumbar spine (L2-4) BMD of less than 0.96 g/cm2 (more than 2 standard deviations below normal peak bone mass). Calcaneal T2' was significantly correlated with calcaneal BMD (r = -0.79, p < 0.0001), BUA (r = -0.59, p = 0.0004) and SOS (r = -0.58, p = 0.0006). T2' was significantly different in postmenopausal women with normal BMD and those with low BMD (p < 0.01). However, the difference was of only borderline significance (p < 0.06) after adjustment for age and years since menopause.