Thiaminase induced thiamine deficiency occurs in fish, humans, livestock and wild animals. A non-radioactive thiaminase assay was described in 2007, but a direct comparison with the radioactive ¹⁴C-thiamine method which has been in use for more than 30 years has not been reported. The objective was to measure thiaminase activity in forage fish (alewife Alosa pseudoharengus, rainbow smelt Osmerus mordax, and slimy sculpin Cottus cognatus) consumed by predators that manifest thiamine deficiency using both methods. Modifications were made to the colorimetric assay to improve repeatability. Modification included a change in assay pH, enhanced sample clean-up, constant assay temperature (37 °C), increase in the concentration of 4-nitrothiophenol (4NTP) and use of a spectrophotometer fitted with a 0.2 cm cell. A strong relationship between the two assays was found for 51 alewife (R² = 0.85), 36 smelt (R² = 0.87) and 20 sculpin (R² = 0.82). Thiaminase activity in the colorimetric assay was about 1000 times higher than activity measured by the radioactive method. Application of the assay to fish species from which no thiaminase activity has previously been reported resulted in no 4NTP thiaminase activity being found in bloater Coregonus hoyi, lake trout Salvelinus namaycusch, steelhead trout Oncorhynchus mykiss or Chinook salmon Oncorhynchus tshawytscha. In species previously reported to contain thiaminase, 4NTP thiaminase activity was measured in bacteria Paenibacillus thiaminolyticus, gizzard shad Dorosoma cepedianum, bracken fern Pteridium aquilinum, quagga mussel Dreissena bugensis and zebra mussels D. polymorpha.