The validity of reference values for vertebral bone mineral density (BMD) reported in the literature has frequently been questioned. The introduction of dual-energy methods, new calibration systems, and automated image evaluation algorithms, has also made it necessary for us to re-evaluate BMD values in a normal population. We present patient selection criteria, CT measurement and evaluation protocols and results of single energy (SEQCT) and dual energy quantitative CT (DEQCT) for trabecular (tBMD) and cortical (cBMD) bone density in 135 males and 139 females. For SEQCT, tBMD loss rates of 1.7 and 1.8 mg/ml per year for males and females, respectively, were determined by linear regression. The respective DEQCT data show smaller decrements with age. Different statistical models for tBMD in females were tested; models based on a multiphase variation of tBMD yielded an improvement of the fit. For cBMD the percent rate of loss with age is much smaller than for tBMD; females showed a significantly higher loss rate than males, however. A comparison of our results with published data is included.