Sixty shellfish samples (mussel [Mytilus edulis], queen scallops [Chlamys opercularis] and king scallops [Pecten maximus]), collected during routine surveillance for paralytic shellfish poisoning (PSP) toxins conducted by the UK Ministry of Agriculture, Fisheries and Food and the Scottish Office Agriculture, Environmental Fisheries Department, were analysed for contamination with PSP toxins by mouse bioassay (MBA), and by a competitive direct enzyme immunoassay (Ridascreen Saxitoxin ELISA). Using the MBA as the reference method (detection limit: 350 micrograms kg-1), no false negative results out of a total of 45 MBA-positive samples and one false positive result out of a total of 15 MBA-negative samples were obtained by ELISA. The correlation coefficient for MBA-positive samples (n = 45) was 0.78. At the regulatory level for PSP toxins (800 micrograms kg-1), 39 of the samples were correctly determined by ELISA to be above or below this level, toxicity was overestimated in two samples, and underestimated in four samples.