Because archived blood specimens are an important but limited resource for conducting epidemiological studies using biomarkers, it is important to develop analytical techniques that minimize the amount of sample needed. We modified an established 1.0-ml blood plasma organochlorine assay to use smaller volumes. We assessed its utility by comparing the accuracy and precision of measurements obtained with different-sized aliquots of spiked plasma from three pools of known concentration of 1,1-dichloro-2,2-bis(p-chlorophenyl)ethylene (DDE) and polychlorinated biphenyls (PCBs; low, medium, and high). There was a modest sacrifice in accuracy using 0.5 as opposed to 1.0 ml. However, the within-batch coefficients of variation, a measure of laboratory error, were consistently low when 0.5-ml aliquots were used. For both DDE and PCB concentrations, this error was less than 5% for the medium and high pools [5-20 parts per billion (ng/ml)] and less than 9% for the low pool (< 1 part per billion). After determining that aliquots of 0.5 ml were sufficient, we performed a blinded quality control analysis of stored plasma. In this study, the within-subject variation was low for DDE and PCBs and substantially lower than the between-subject variation, suggesting that the assay would rank subjects with reasonable precision. Our results suggest that use of 0.5-ml as opposed to 1.0-ml aliquots should not compromise the power of a nested case-control study to detect differences between subjects and would thus save plasma for future research. For populations with very low levels of organochlorines, however, the larger volumes should still be used.