In the control of larval sea lamprey (Petromyzon marinus ) with 3-trifluoromethyl-4-nitrophenol (TFM) in tributaries of the Great Lakes, occasional kills of other fishes have caused concern about the effects of the chemical on non-target organisms. Stream treatment rates have been based on previous application rates, alkalinity measurements, results of on-site toxicity tests, or combinations of these. Laboratory studies in 1987 showed that pH is the primary factor that affects the toxicity of TFM (the lower the pH, the greater the toxicity): even small changes in pH alter the toxicity, whereas substantial changes in alkalinity have little effect. In 12-h exposures, the 96-h LC50 for TFM to rainbow trout (Salmo gairdneri ) ranged from about 0.9 mg/L at pH 6.5 to > 100 mg/L at pH 9.5, but (at pH 7.5) the LC50's differed little at total alkalinities of about 18 mg/L and 207 mg/L. Decreases in pH as small as 0.5 pH unit caused nontoxic solutions to become toxic to rainbow trout. Some kills of non-target fish during stream treatments were reportedly caused by decreases in pH, and (conversely) that some stream treatments for sea lampreys were ineffective because pH increased.