Bacteriolysis was found to be markedly dependent on ionic strength and pH, with optima at µ = 0.06 and pH 8.3–8.5, respectively. The method of cultivating the cells also influenced the rapidity of lysis.

Lysis was temperature-dependent and was exhibited by all samples of human serum tested. Microscopically, organisms incubated with serum were observed to swell, loose their rod shape and eventually burst, leaving remnants of the cell membrane in suspension. Sphaeroplasts were obtained by brief exposure of cells to serum followed by dilution into 5 per cent sucrose.

The bacteriolytic reaction was shown to require complement. No definite requirement for properdin or specific antibody in this system could be demonstrated by the absorption of serum with zymosan and with homologous cells respectively. The latter procedure was found to reduce bacteriolytic activity by removal of serum lysozyme. Absorption of serum with bentonite also led to loss of bacteriolytic activity which could be restored with lysozyme. The organism was not lysed by lysozyme alone, but lysis occurred with lysozyme + EDTA in tris buffer.

The possibility of complement acting independently of antibody or properdin, in certain instances, is discussed in relation to bacterial cell wall structure. Data are presented supporting the hypothesis that the "substrate" of complement in cell membranes is a lipid or lipoprotein.