T. M. SONNEBORN 183 



varieties from the other three more remotely related ones (i.e., 

 varieties 1, 3, and 9). First, in varieties 2, 4, and 8 each strain 

 usually possesses several alternative serotypes that can be stably 

 maintained under appropriate conditions. Second, so far as now 

 known, corresponding serotypes of different strains of the same 

 variety seldom if ever totally fail to cross-react to each other's 

 homologous antiserum. Third, in any one strain, usually more 

 than one stable serotype may be expressed in different sublines 

 under the same conditions; that is, there is a strong cytoplasmic 

 element controlling the persistence of a given serotype. Fourth, 

 corresponding serotypes of different strains may require very dif- 

 ferent conditions for their expression. 



In all four of these respects the varieties 1, 3, and 9 of the 

 other group seem to be like each other but different from va- 

 rieties 2, 4, and 8. The phenomena have been most fully worked 

 out in variety 1 (Beale, 1954). Although each strain of variety 

 1 has a considerable array of potentially producible serotypes, 

 only three serotypes, D, G, and S, are very stable, and these are 

 regularly found in all natural strains. The specificities of any 

 one of these serotypes in different strains may be much alike, 

 but often the differences in specificity are so great that no cross 

 reactions occur and correspondence is detectable only by genetic 

 and physiological means. Further, the expression of these three 

 serotypes is in a uniform way in all strains closely dependent 

 upon temperature. Thus, in any strain serotype S requires for 

 its expression lower temperature, and serotype D higher tempera- 

 ture, than serotype G. This results regularly in the expression of 

 only one stable serotype in any one strain under given conditions 

 and evidence of a cytoplasmic element in the persistence of a 

 serotype is as a rule inconspicuous or lacking. In so far as evi- 

 dence is available, these features of the serotype system appear 

 also to be operating in varieties 3 and 9. 



The differences among varieties 1, 3, and 9 concern mainly 

 the array of serotypes in each and the specificities of correspond- 

 ing serotypes when such are found. These differences are com- 

 parable to those found among varieties 2, 4, and 8. The two varie- 

 ties 1 and 3 (which are more closely related than either is to 



