The Massachusetts Department of Public Health, with the assistance of the Centers for Disease Control, conducted a study to determine the prevalence of elevated levels (> 30 ng/ml) of polychlorinated biphenyls (PCBs) in serum taken from residents of the greater New Bedford area in Massachusetts. The criteria and procedures used to establish interlaboratory comparability for measurements for PCBs in human serum, the establishment and performance of the quality control system and the comparability of results on human serum samples from the New Bedford study are described. Aspects of interlaboratory comparability addressed include the establishment of common extraction and analytical methods, joint analyses of bovine serum pools (both in vitro-spiked pools and in vivo pools from a cow that was fed PCBs), establishment of quality control charts and rules for acceptability of analytical runs and joint analyses of a subset (n = 207) of the human serum samples from the New Bedford study. The 207 jointly analyzed samples had PCB levels that ranged from 1 to 214 ng/ml and had an interlaboratory correlation coefficient of 0.96.