When studying the behavior (carry, dispersion, persistence) of Bacillus thuringiensis var. israelensis (B.t.i.) formulations used in the treatment of rivers or streams for black fly control, a large number of samples containing small quantities of B.t.i. are required for proper analysis. Freezing is a useful procedure to prevent enzymatic alteration or bacterial growth in samples before bioassays are to be performed. Using Aedes atropalpus neonate larvae, we studied the effect of freezing and thawing of B.t.i. aqueous suspensions by looking at mortality response parameters such as the slope and the LC50 of the probit regression. Initial concentration values of 1, 5, 10 and 20 mg/liter at the moment of freezing of the B.t.i. suspensions did not significantly affect toxicity. The number of freeze-thaw cycles greatly increased the LC50 values without much change to the slope of the log-probit regressions. We derived an equation that allowed us to compensate for the loss of toxicity of a given B.t.i. sample, knowing the number of freeze-thaw cycles.