In order to evaluate cast methods for estimating age from the os pubis, 202 female and 116 male pairs of pubes collected at autopsy were blindly evaluated for age using the McKern-Stewart or Gilbert-McKern and Suchey-Brooks methods. Performance of the methods was measured by mean absolute deviation of true age from scored interval means and frequencies of true age falling within established standard deviations from the mean. By the standards of evaluation used in this study, all aging methods based on the os pubis proved disappointing with regard to both accuracy and precision. Deviations tended to be lower for younger individuals and higher for the aged. The results suggest that the Suchey-Brooks methods are superior in forensic applications and that the racially specific refinement for males should be used. Age estimates should include plus or minus 2 standard deviations, and chances of error should be considered especially when advanced age or debilitation is suspected. 9 references, 2 tables, and 1 figure (Author abstract modified)