40() LOCAL TLSSIIE REACTIVITY 



and ill miiliij)Ic pi ojjort ions. Ilic effects ol endotoxins npon test 

 animals were sliown by Pleitier and Bcssau, and more recently 

 by Minray (h)2()) to be ine,<;ular in their effect. 



The interesting lesnhs obtained with the serum of Gory and 

 Grasset depended ii])on the protection of guinea pigs against sub- 

 ctitaneous infections with H\e B. typhosus. Unfortunately, no ac- 

 cinate determination oi the j)otency of the serum is possible 

 imder the experimental (onditions described ])y these authors 

 (1928, 19.^^0) . 



The work ot Felix (i9'55, i9-^fi) presently under way deserves 

 special consideration. This author apparently obtained encourag- 

 ing results in treatment of tyj)h()id fever by the use of an immune 

 serum (ontaining \'i and () antibodies. The Vi antibody confers 

 protection against inlection of mice with virident strains of B. ty- 

 phosus and the () antibody seems to be anti-endotoxic. 



Preparation cf phenomenon-neutralizing anti-typhoid horse 

 sera: 



In the ^vork about to be described attempts ^vere made to 

 develop an anti-typhoid serum of a standardized phenomenon- 

 neutralizing potency for use in patients. Horses were immimized 

 by a mixture of active principles differing in their neutralizability 

 as Avell as active principles derived from smooth and rough vari- 

 ants previously described (p. 101). 



The inclusion of active principles derived from rough variants 

 seems to be desirable inasmuch as bacterial variation of this sort 

 may also be presumed to occur hi vivo in the course of an infec- 

 tion of long chnation. Shotdd this change occur in vivo, there 

 may appear in the patient's body rough reacting factors, non- 

 neutralizable by the antibodies of the strain responsible for the 

 initial infection, since the rough variants Avere shown here to 

 produce active principles of ne^v antigenicity (Shwartzman, 

 1930^) . Possibly the formation of rough strains may be held 

 responsible for relapses of typhoid fever. Until now, the toxicity 

 of a new antigenic specificity of bacterial filtrates derived from 

 rough B. typhosus strains has not been recognized. In this con- 

 nection it is of interest to mention the w^ork of Durham (1901) , 

 Avho many years ago assumed that the original typhoid infection 

 is produced by various strains of allied bacilli. According to his 

 view, if the different strains are similar, antibodies are produced 

 to the several strains and no relapses occur. If a particular strain 

 predominates in the primary attack, though effective antibodies 



