K. LANDSTEINER 



893 



tions proposed by Jansky and by Moss.' For that reason several authors prefer a no- 

 menclature by letters as employed in the present article. This designation of the 

 four groups is O, A, B, AB, which, as will be seen, represents the constitution of the 

 corpuscles with respect to isoagglutinogens, the letter O indicating the absence of 

 both agglu finable substances A and B. 



The correlation of the various nomenclatures and the distribution of the agglu- 

 tinogens and agglutinins among the groups are shown in Table I. The scheme brings 

 out the rule, which is not self-evident, that the serum regularly contains the aggluti- 

 nins active for the absent agglutinogens. The other regularity, that corresponding 

 agglutinins and agglutinogens do not coexist in one blood, is plausible a priori. Conse- 

 quently the characteristics of the groups are as follows: Group 0: The corpuscles 



TABLE I 



TABLE II 



are not agglutinated by any serum; the serum agglutinates the corpuscles of all other 

 groups (cells A generally more intensely than cells B). Group A: The corpuscles are 

 agglutinated by the sera of groups and B; the serum agglutinates the corpuscles of 

 groups B and AB, Group B: The corpuscles are agglutinated by the sera of groups 

 O and A; the serum agglutinates the corpuscles of groups A and AB. Group AB: 

 The corpuscles are agglutinated by the sera of all other groups; the serum does not 

 agglutinate the corpuscles of any group. These statements are summarized in Table 

 II where-f- denotes agglutination, — no agglutination. 



It is evident, then, that an unknown blood can easily be grouped by employing 

 either two sera or two sorts of corpuscles known to belong to groups A and B or serum 

 and corpuscles of either group A or B. The technique will be discussed below. The 

 groups can be determined also without the use of standard sera or bloods by cross- 

 testing a sufficient number of blood samples. In this way groups O and AB are recog- 

 nized immediately; but the differentiation of groups A and B is only possible by tak- 

 ing into consideration that group A is much more frequent than B among white races. 



' See J.A.M.A., 76, 130. 1921. 



