Determining dissolved concentrations of polychlorinated biphenyls (PCBs) in sediment pore (interstitial) water with conventional solvent extraction methods is problematic because obtaining large (liter) quantities of pore water, separating it from the sediment, and removing the influence of colloids can be very difficult. However, solid-phase-microextraction (SPME) can achieve similar detection limits using milliliter water samples as achieved with organic solvent extraction requiring a liter of pore water. Five different SPME sorbents were evaluated for their ability to yield the best detection limits for di- to octachlorobiphenyl congeners, both with GC/ECD and with GC/MS (both positive ion EI and negative ion CI). SPME using the 7 microm PDMS fiber with GC/MS (positive ion EI) yielded the best combination of signal-to-noise and selectivity using a 30 min extraction, although ECD was also suitable. Pore water was obtained by centrifuging wet sediment followed by flocculation to remove colloids. Quantitative calibration was simplified by adding dichloro- to hexachlorobiphenyl internal standards chosen to be compatible with either ECD or MS detection. Calibration curves and relative response factors (including the SPME and GC steps) were determined for all 62 PCB congeners that are present in above-trace quantities in commercial Aroclors. Calibrations were linear (r(2) typically >0.995) from low pg/mL to ng/mL concentrations, with near zero intercepts. Detection limits for all individual PCB congeners ranged from <1 to 3 pg/mL using 1.5 mL water samples. Dissolved organic matter (DOM) had no measurable effect on dichloro- and trichlorobiphenyls, but did contain about 10 to 25% of the tetrachlorobiphenyls and up to 60% of the hexachlorobiphenyl congeners. Log DOC/water partitioning coefficients (log K(DOC)) ranged from 3.6 to 4.6 for 2,3,5,6-tetrachlorobiphenyl and from 4.2 to 5.5 for 3,3',4,4',5,5'-hexachlorobiphenyl.