Determination of PCBs in biological tissue extracts by enzyme-linked immunosorbent assays (ELISAs) can be problematic, since the hydrophobic solvents used for their extraction and isolation from interfering biochemicals have limited compatibility with the polar solvents (e.g. methanol/water) and the immunochemical reagents used in ELISA. Our studies of these solvent effects indicate that significant errors can occur when microliter volumes of PCB containing extracts, in hydrophobic solvents, are diluted directly into methanol/water diluents. Errors include low recovery and excess variability among sub-samples taken from the same sample dilution. These errors are associated with inhomogeneity of the dilution, which is readily visualized by the use of a hydrophobic dye, Solvent Blue 35. Solvent Blue 35 is also used to visualize the evaporative removal of hydrophobic solvent and the dissolution of the resulting PCB/dye residue by pure methanol and 50% (v/v) methanol/water, typical ELISA diluents. Evaporative removal of isooctane by an ambient temperature nitrogen purge with subsequent dissolution in 100% methanol gives near quantitative recovery of model PCB congeners. We also compare concentrations of total PCBs from ELISA (ePCB) to their corresponding concentrations determined from capillary gas chromatography (GC) in selected fish sample extracts and dialysates of semipermeable membrane device (SPMD) passive samplers using an optimized solvent exchange procedure. Based on Aroclor 1254 calibrations, ePCBs (ng/mL) determined in fish extracts are positively correlated with total PCB concentrations (ng/mL) determined by GC: ePCB = 1.16 * total-cPCB - 5.92. Measured ePCBs (ng/3 SPMDs) were also positively correlated (r² = 0.999) with PCB totals (ng/3 SPMDs) measured by GC for dialysates of SPMDs: ePCB = 1.52 * total PCB - 212. Therefore, this ELISA system for PCBs can be a rapid alternative to traditional GC analyses for determination of PCBs in extracts of biota or in SPMD dialysates.