We evaluated the precision and accuracy of a peripheral quantitative computed tomography (pQCT) device, Densiscan-1000 (Scanco Medical AG Bassersdorf, Switzerland). Seven forearms of cadaveric specimens were scanned on the pQCT device using the procedure recommended by the manufacturer. Bone mineral density (BMD) was determined in the distal radius and in the more proximal, largely cortical site. Cortical area was assessed on tomograms and mean cortical width calculated assuming a circular ring model. The procedure was repeated five times either with repositioning or without repositioning to obtain the short-term precision error. The coefficients of variation (CV) of BMD at the distal and the proximal site were 0.32% (90% CI = 0.26–0.41) and 0.20% (0.16–0.26), respectively. The CV of cortical width was 1.76% (1.45–2.26). Both BMD and cortical width were not significantly influenced by repositioning. After these measurements, we evaluated the effects of soft tissue on the BMD determination. The BMD was slightly underestimated by removing the surrounding soft tissue, and the difference was 1.4% on average. Subsequently the bone segments corresponding to the measurements were dissected. Cortical area and cortical width estimated from tomograms were compared with those from the surface images of dissected bone sections. Cortical width above 1 mm could be accurately estimated. Finally, the bone segments were ashed to evaluate the accuracy of bone mineral content (BMC) determination. Ash weight and BMC were highly correlated (r = 0.976) with an error of 10.75%. In summary, pQCT using Densiscan-1000 appears to be able to be assess both the BMD and the cortical width at the radius with high precision and moderate accuracy.