We draw the following conclusions from our experiments on the antigenic properties of chemically pure casein and some of its split products.

Casein and paranuclein have distinct antigenic properties, particularly as shown by their ability to sensitize guinea pigs for subsequent anaphylactic intoxication by each other or by milk. This sensitizing ability and a corresponding ability to intoxicate are indistinguishably equivalent, under the conditions employed. On immunizing rabbits by repeated injections of paranuclein or of casein, and subsequently testing their sera for precipitins and fixation antibodies, it was found that casein apparently produces them much more readily, giving an antiserum that reacted (fixation) in very high dilution with casein (0.000,000,1 of a 1 per cent. solution), but much less strongly with paranuclein. Only one of two antiparanuclein sera showed the presence of antibodies to paranuclein by the delicate fixation reaction, and that in relatively small amounts. The two antibodies to casein and to paranuclein are, in the case of casein quantitatively, and in the case of paranuclein absolutely specific.

A solution of the products of complete peptic digestion of casein fails to sensitize to paranuclein and gives no fixation reaction with an anticasein or antiparanuclein serum. It intoxicates animals sensitized to paranuclein but no more markedly than it does normal animals. It also fails to show specific intoxication in an animal that has been sensitized by the same substance.

The amino acids, glutamic acid, and leucin, the principal components of their kind in casein, and in the same proportion therein present, likewise failed to show antigenic properties. They do not sensitize animals to milk intoxication or to intoxication by themselves, and likewise failed to produce precipitins in rabbits in a preliminary experiment.

These experiments are regarded as a fairly systematic analysis of the antigenic properties of split products of a single protein. They are analogous to, though less complete than the work of Wells (6) on egg-white. They seem to present the additional advantage of dealing with what is probably the only protein certainly known chemically, and in its purest form. They serve, moreover, as an introduction to the following study of the antigenic properties of a combined protein.

This content is only available as a PDF.