This work was aimed at advancing knowledge of the heat-labile toxin (HLT) of Bordetella pertussis along several lines, ultimately directed at gaining a better understanding of the possible roles of the toxin in the pathogenesis of B. pertussis infections and in the development of immunity. The following observations were made: 1 Assay and stability of the toxin 1.1 When assaying the lethal toxicity of HLT in the mouse, neither the sex nor the weight of the animal had a significant influence on the result. 1.2 The assay of HLT by its haemorrhagic activity in 4 d old mice was at least 10 times as sensitive as the LD50 estimation in 3-4 wk old mice. 1.3 Contrary to the report of Violle, Adult mice (strain Hr/Hr) were found to be susceptible to the dermonecrotic activity of HLT. 1.4 HLT was not haemolytic for sheep, rabbit, horse, human or cod red cells, even in doses as large as 2.5 mouse LD50 per ml. There was no hot-cold lysis. 1.5 Circumstantial evidence is presented that HLT may exist in an inactive form which can be converted by proteases into an active form. 1.6 For storing HLT, freezing (-20°C) was found to be satisfactory. The mean rate of inactivation of 4 preparations tested over periods up to 3 yr was 10% per annum.