1. A group of 41 non-serum treated patients with Type I, II or III pneumococcus pneumonia were studied during their disease and convalescence with respect to their skin reactions to specific pneumococcus polysaccharides and, in most instances, for the presence of circulating agglutinins and protective antibodies for all these 3 types.
2. One-half of the Type I and two-thirds of the Type II and Type III recovered cases gave the typical immediate "wheal and erythema" response to the homologous polysaccharide at or about the time of recovery. All cases tested showed protective antibodies and almost all showed agglutinins for the homologous pneumococcus. In the fatal cases, in general, positive cutaneous reactions and circulating antibodies were not obtained.
3. In cases of pneumonia receiving repeated cutaneous inoculations with various types of specific polysaccharide, antibodies for pneumococci differing from the infecting type but corresponding to the types of carbohydrate injected were present 1 week or later after such injections. These heterologous antibodies were most frequently demonstrated for Type II and were probably the result of immunization by means of the cutaneous injections.
4. Positive skin responses to homologous polysaccharides and corresponding circulating antibodies were demonstrated with similar frequency in the first 3 weeks after crisis in patients who had not previously received intracutaneous injections. In such patients heterologous antibodies were rarely found.
5. Typical skin reactions with the specific pneumococcus polysaccharides and mouse protective antibodies were demonstrated independently in a number of hospital patients who had had no recent history of pneumonia.
6. Some patients with demonstrable foci of persistent infection or with latent infections which later proved fatal showed positive cutaneous responses to the homologous type polysaccharide and circulating specific antibodies for the corresponding type.
7. The agglutination test, though less sensitive than the mouse protection test for determining the presence of antibody, has many advantages over the latter and is simplest to use in following the course of the untreated pneumonia.