Certain interesting facts have been obtained regarding the action of normal swine serum, but in the light of present knowledge it is not easy to interpret them. It has been found that the normal serum possesses the property of protecting mice from infection with virulent pneumococci, and of agglutinating virulent S pneumococi and also avirulent R pneumococci. The protective action of the serum is specific, that is, absorption of the serum with pneumococci of one type removes or destroys only the property of the serum which is responsible for its protective action against pneumococci of that particular type, leaving the serum still active against pneumococci of other types. In this particular, therefore, the swine serum resembles a polyvalent antipneumococcus serum produced by artificial immunization. In other particulars, however, the swine serum differs from that produced byartificialimmunization. Inthefirstplace, theprotectiveaction exhibited by normal swine serum against type-specific pneumococci is very slight compared with that of animals which have been artificially immunized, even though the latter have received only a very few injections of the specific antigen. Moreover, the protective properties of the swine serum are destroyed by heating to 65 degrees C., and they disappear after a few weeks, even if the serum is kept in the cold. Also the protective action of the serum is diminished after the serum has been mixed with non-type-specific avirulent R pneumococci and the latter have been removed by centrifugalization. A further difference between the action of swine serum and that of the antipneumococcus serum produced by artificial immunization is that although the swine serum is type-specific in its action in protecting mice, nevertheless when this serum is mixed with a purified type-specific polysaccharide, neither precipitation nor fixation of complement results; and, finally, the action of the serum in protecting mice against pneumococci of a particular type is not inhibited by the addition of the homologous polysaccharide to the serum. These latter facts relating to the specific polysaccharide, and especially the fact that the specific protective action disappears after applying the so called absorption technique with non-type-specific organisms, are most difficult to harmonize with the present conceptions of the nature of pneumococcus type specificity. It is not proposed at the present time to offer any theoretical explanation of these phenomena. The studies have also confirmed the observations of Robertson and Sia, that, if a proper technique be employed, normal swine serum exhibits the property of agglutinating type-specific pneumococci, and that the agglutinins for pneumococci of one type may be specifically absorbed, leaving those for pneumococci of other types unchanged. Finally, it has been shown that swine serum causes the agglutination of avirulent R pneumococci. This property of the serum is not destroyed by heating at 65 degrees C., and therefore probably depends on factors other than those responsible for the agglutination of S pneumococci, and for the protective action of the serum in mice.