A rapid liquid chromatographic (LC) method with postcolumn oxidation and fluorescence detection (excitation 330 nm, emission 390 nm) for the determination of paralytic shellfish toxins (PSTs) in shellfish tissue has been developed. Extracts prepared for mouse bioassay (MBA) were treated with trichloroacetic acid to precipitate protein, centrifuged, and pH-adjusted for LC analysis. Saxitoxin (STX), neoSTX (NEO), decarbamoylSTX (dcSTX), and the gonyautoxins, GTX1, GTX2, GTX3, GTX4, GTX5, dcGTX2, and dcGTX3, were separated on a polar-linked alkyl reversed-phase column using a step gradient elution; the N-sulfocarbamoyl GTXs, C1, C2, C3, and C4, were determined on a C-8 reversed-phase column in the isocratic mode. Relative toxicities were used to determine STX-dihydrochloride salt (diHCl) equivalents (STXeq). Calibration graphs were linear for all toxins studied with STX showing a correlation coefficient of 0.999 and linearity between 0.18 and 5.9 ng STX-diHCI injected (equivalent to 3.9-128 microg STXeq/100 g in tissue). Detection limits for individual toxins ranged from 0.07 microg STXeq/100 g for C1 and C3 to 4.1 microg STXeq/100 g for GTX1. Spike recoveries ranged from 76 to 112% in mussel tissue. The relative standard deviation (RSD) of repeated injections of GTX and STX working standard solutions was < 4%. Uncertainty of measurement at a level of 195 microg STXeq/100 g was 9%, and within-laboratory reproducibility expressed as RSD was 4.6% using the same material. Repeatability of a 65 microg STXeq/100 g sample was 3.0% RSD. Seventy-three samples were analyzed by the new postcolumn method and both AOAC Official Methods for PST determination: the MBA (y = 1.22x + 13.99, r2 = 0.86) and the precolumn LC oxidation method of Lawrence (y = 2.06x + 12.21, r2 = 0.82).