The effect of heating on plasmin activity in various media, including phosphate buffer pH 7·0, skim milk, blood plasma, solutions of casein and solutions of whey proteins were investigated. Plots of log residual activity υ. heating time were linear at all temperatures from 63 to 143 °C. In buffer solutions the presence of casein led to substantial substrate protection, the Arrhenius plots being linear both in the presence and absence of casein. The activation energy, Ea, for the inactivation reaction, was 62·4 kJ/mol in buffer alone and 58·4 kJ/mol with casein present at 25 mg/ml. In skim milk, despite the presence of casein at a similar concentration, plasmin was no more stable to heat than in buffer alone, and a curved Arrhenius plot was obtained indicating a more complex inactivation mechanism. Heating in the presence of proteins having free -SH groups accelerated the inactivation of plasmin. The role of -SH groups was confirmed by experiments with added α-lactalbumin, in which no free -SH groups occur, and reduced carboxymethylated β-lactoglobulin, both of which were without effect. In blood plasma, plasmin was less stable to heat than in buffer (pH 7·0) or in skim milk. Plasminogen behaved very similarly to plasmin either when activated to plasmin with urokinase before heating or when activated afterwards. A hypothesis is presented to describe the heat inactivation and denaturation of plasmin. Technologically important findings are that in skim milk plasmin was largely unaffected by pasteurization conditions and 30–40% of its activity remained even after ultra high temperature processing conditions.