The major aim of this study was to develop an accurate method of age prediction for a wide range of ages, based on the roentgenographic assessment of the hand bones, for use in paleoanthropology and forensic medicine. The roentgenographs of the hand bones were assessed in 5756 individuals, 2683 males and 3073 females, aged 17-93 and 17-89, respectively, belonging to 9 ethnic groups from the former USSR, Israel, and Sinai peninsula (Egypt). For each roentgenogram an equidistant osteographic score (OSS) including the descriptive criteria of bone age, such as (1) osteophytes or nodes of Heberden in the periarticular regions and at sites of tendon attachment, (2) manifestations of osteoporosis, (3) signs of sclerosis, and (4) non-traumatic articular deformities, was estimated. Regression analysis, that included linear, non-linear, logistic, and stochastic models, was used to evaluate the relationships between age and OSS. This relationship was best described by a logistic regression function. Results of the logistic regression analysis clearly indicate that OSS is a strong predictor of an individual's age, with r2 values ranging from 0.671 to 0.901 (p < 0.001). Our standard errors of estimate were ranging from +/- 4.2 to +/- 7.3 years, comparing favorably with most known methods of bone age assessment. This study provides an efficient method of age prediction, which allows to extend the upper limit of prediction to the age of 70 years with acceptable accuracy.