The chemical and immunological characteristics of the type-specific substance (M) of Streptococcus haemolyticus are considered. 1. A summary of the evidence for the protein nature of this substance follows: (a) It is precipiated by the usual protein precipitants such as, dilute alcohol, dilute acetic acid, and picric acid. (b) It contains 14 per cent protein nitrogen after considerable purification. (c) It is progressively destroyed by removal of the NH(2) group by treatment with nitrous acid. (d) It is completely and readily digested by trypsin and by pepsin. 2. "Purified" extracts react in relatively high dilution with homologous antibacterial sera, but do not precipitate most heterologous antibacterial sera or sera potent in non-type-specific antibodies for the group reactive nucleoprotein P or for the species-specific probable carbohydrate C. Attempts to immunize rabbits with the type-specific protein have been unsuccessful, with simple salt solution extracts of streptococci as well as with purified solutions. This protein seems, therefore, to have the characteristics of a haptene. The type-specific substance (M) is contrasted with the so called nucleoprotein (P) which shows group relationships with nucleoproteins of related species and is the only fraction of hemolytic streptococcus extracts so far obtained which, after separation from the bacterial cell, is a true antigen leading to antibody production when injected into rabbits. The occurrence of another non-type-specific protein (Y) is suggested by occasional cross-reactions of purified M with certain antibacterial sera. Since it has not been separated from extracts containing the type-specific M, little is known of it either chemically or serologically. The cross-reaction disappears on tryptic or peptic digestion of the extract. The fact that such extracts do not show cross-reactions with anti-P sera is evidence that this non-type-specific protein is not P.