1. Normal guinea pigs treated by four to six instillations of horse serum into the nose on alternate days become either hypersensitive or refractory to an intravenous injection of 0.38 cc. of serum given 16 days after the last instillation. If the amount of serum in each instillation is as much as 0.2 cc., anaphylactic death is caused by the toxic injection. If the amount of serum in each instillation is reduced to 0.04 cc. the first intravenous injection is without marked effect, and a second injection and subsequent injections of the same amount of antigen are well tolerated in about half the cases.
2. The effect produced by a given dose of serum, whether protective or anaphylactic, depends probably upon the extent of contact with the mucous membrane of the nose.
3. Guinea pigs which, after nasal treatment, have become tolerant to a definite maximum intravenous injection of the antigen appear to increase the degree of their tolerance, at least up to a resting period of more than 4 months. The same does not hold in animals immunized by the peritoneal route.
4. The first two or three instillations of a series probably determine the biologic character, whether of hypersensitiveness or hyposensitiveness, of reaction towards the serum.
5. It is probable that, contrary to the case in parenteral sensitization, hypersensitiveness and protection, respectively, set up by nasal instillations and not followed by parenteral injections, gradually disappear in about 50 to 100 days.
6. We have failed in attempts to eliminate hypersensitiveness, due to subcutaneous injection of serum, by nasal instillations which would protect the normal animal from the development of anaphylaxis.
7. It is suggested that the principles of prophylaxis evolved under these relatively simple conditions should be applied in the study of infectious disease.