The serological properties of six new pneumococcal capsular types are described. A table listing all 90 pneumococcal types and their cross-reactions is included. When Streptococcus pneumoniae type 47A was described in 1972 (7), the number of known pneumococcal capsular types totaled 83 (see also reference 6). In 1985, Austrian and coworkers described a new type: 16A (1). In 1979, a worldwide pneumococcal typing surveillance study, aimed mainly at typing isolates from blood and cerebrospinal fluid, was initiated under the auspices of the World Health Organization. Since then, more than 25,000 strains, most from normally sterile body sites, have been typed in Copenhagen, Denmark (2, 9). Typing was performed according to a procedure described previously (6). During the course of this study, six new types, 10B, 10C, 11D, 12B, 25A, and 33D, with which this report deals, were detected by the procedure described by Mørch (8) based on the dictum: ''the serological identity of two types can only be proved by means of crossadsorption of the mutual immune sera'' (4). Antisera to these strains were prepared in two or more rabbits according to the standard immunization procedures of the World Health Organization Reference Centre (6), and antibodies to them are now included in Omniserum, in the diagnostic serum pools D, E, and I, and in antisera to groups 10, 11, 12, 25 (previously type 25), and 33, made at Statens Seruminstitut, Copenhagen, Denmark (3, 6). Antigenic formulas (4), an expression used throughout this paper, represent arbitrary designations of cross-reactions as seen by the capular reaction (4-6). Mørch (8, p. 98) wrote: ''Antigen a signifies the factor characteristic of the individual type or common to the types of a group. The letters b, c, d and so on indicate the additional partial antigens, which in certain cases may have developed more strongly than the antigen particular to the type''. For instance, the two factor sera, 6b and 6c, necessary to distinguish type 6A from type 6B, are of course made by reciprocal absorption which takes experience because there is only a relatively narrow interval between incomplete absorption of the unwanted reaction and complete absorption of all type-specific antibodies, including heterologous antibodies. The reason for including Tables 1 to 5, therefore, is that the antigenic similarities and differences shown represent the only way of demonstrating that the six new types are, in fact, different from the ones to which they are most closely related.