The efficacy of detecting asbestos (1332214) related thoracic anomalies by the conventional posterior anterior (PA) and combined PA roentgenogram (PAR) procedures was studied. Asbestos workers (555 subjects) exposed for less than 15 to more than 30 years were examined twice by PA and PAR. Error and bias were controlled by comparing intra reader, inter reader, and inter method agreement values and by averaging the readings obtained. The types of anomalies examined were pleural thickenings in one or both lungs, pleural plaques noted in one or both readings, and profusion of small opacities, their size, shape, and numbers. The detection rate for pleural thickening for example by PA and PAR rose from 9.3 and 4.7 percent found at less than 15 years of exposure to 19.5 and 26.2 percent for 30 or more years of exposure for the two methods, respectively. The total number of cases involved at these two time periods were 57 and 547, respectively. The use of PAR produced a significantly higher rate of detection of thoracic anomalies than PA. Thus, 22, 103, and 93 percent higher rates were found for the pleural thickening, pleural plaque, and small opaque profusion, respectively, with PAR than PA. Intra reader values were inconsistent indicating that use of PAR did not uniformly reduce variability in interpretation over that obtained by PA alone. Inter reader values differed by 30 and 50 percent depending on the type of thoracic abnormality assessed. There was also a lack of agreement between reading procedures and the standards as perceived by the readers. The authors conclude that although PAR has a higher rate of detection for asbestos related chest abnormalities than PA, it is not a reliable procedure for use in screening programs.