K. LANDSTEINER 



899 



change was brought about by the action of a special bacillus which makes the blood 

 cells sensitive to an agglutinating agent present in human serum. Thomsen's phe- 

 nomenon is to be kept in mind on testing blood specimens that are not perfectly 

 fresh. 



The distinguishing features of pseudo-, auto-, and isoagglutination are represented 

 in Table III. 



DEFICIENT REACTIONS 



Among the atypical cases reported there are some in which the agglutinins ex- 

 pected from the rule of the groups were deficient. Although these findings were cer- 

 tainly due in a number of cases to overlooking weak reactions, it is possible that in 

 rare instances there is a temporary or permanent deficiency of agglutinins. Grove, 

 however, did not find such an instance among over two thousand examinations. 



TABLE III 



Absorption of the active prin- 

 ciple 



Influence of temperature 

 \' ariations 



Effect of dilution of the 

 ?erum 



PseudoagKlutination 



Not absorbable 



Not less, rather more, 

 active at 37° C. 

 than at lower tem- 

 perature 



Disappears on slight 

 f'ilution 



Autoagglutination 



Absorbable 



Generally occurring 

 only at low temper- 

 ature 



Stands moderate dilu- 

 tion 



Isoagglutination 



Absorbable 



Not much affected by 

 temperature varia- 

 tions between 0° 

 and 37° 



Stands moderate dilu- 

 tions 



SUBGROUPS OF GROUP A 



It was demonstrated for the first time by von Dungern and Hirschfeld' that after 

 absorbing a group B serum with individual group A corpuscles the supernatant fluid 

 reacted on some and not on other red cells of group A. This experiment, indicating a 

 subdivision of group A, has been confirmed by several authors.^ According to Coca, 

 the non-reacting corpuscles amount to about 25 per cent of all samples of group A. 

 Mino and Lattes and Cavazutti attributed these observations to purely quantitative 

 differences in the agglutinability of the corpuscles. This view appeared to be sub- 

 stantiated by their finding that all anti-A agglutinins can be entirely removed from 

 a serum by treating it with the non-reacting corpuscles, particularly at low temper- 

 atures, and that these corpuscles are less sensitive toward sera and B than the others. 

 Evidently, then, the assumption seemed warranted that the less sensitive corpuscles 

 would not react with the serum when the amount of agglutinins was diminished by 

 incomplete absorption while the more sensitive cells would still be acted upon. It was 

 possible, however, to separate two fractions of agglutinins from sera O and B, the one 

 acting on both types of group-A cells, the other chiefly on one of them.^ This leads 

 to the conclusion that the corpuscles are qualitatively different. One may consider 

 either two slightly different agglutinogens A or else an additional agglutinogen in the 



'von Dungern, E., and Hirschfeld, L.: Zischr. f. Immiinildlsforsch. u. exper. TJierap., 13, 527. 

 1911. 



2 Schiitze, H.: Brit. J. Expcr. Path., 2, 26. 1921; Coca, A. F., and Klein, H.: /. Immunol., 13, 

 477. 1923; Guthrie, C. G., and Huck, J. G.: Bull. Johns Hopkins Hosp., 34, 37, 80, 128. 1923. 



3 Landsteiner, K., and Witt, D. H.: Joe. cit. 



