ENDOTOXIN OF WHOOPING-COUGH BACILLUS. 489 



of all be separated in order to produce its antibody. In previous 

 articles we have insisted on the fact that the rapid death of guinea- 

 pigs that have received an injection of a sufficient number of 

 living whooping-cough bacilli in the peritoneal cavity is due to an 

 intoxication from the poison that comes from the bacteria, and 

 not to an infection, properly speaking. The bacteria injected do 

 not increase to any extent in vivo, and do not invade the animal 

 body; they are retained at the point of inoculation, and are fre- 

 quently phagocyted. At autopsy, however, considerable lesions 

 are found and in particular a marked peritoneal exudate of hemor- 

 rhagic character, abundant disseminated ecchymoses, and pleural 

 effusions which are so extensive as to produce asphyxia, and are 

 the immediate cause of death. All these phenomena may be pro- 

 duced by almost perfectly limpid extracts containing endotoxin 

 which we have obtained by employing . Besredka's method as 

 described for typhoid, plague and dysentery endotoxins. We 

 inoculate the whooping-cough bacillus on our blood-agar medium, 

 and in making this medium we prefer horse blood in order to obtain 

 large amounts of media readily; we leave the cultures in the incu- 

 bator for 3 days, and then remove the layer of bacteria with a 

 glass rod and suspend them in a small amount of salt solution. The 

 very thick emulsion obtained in this way is dried in a vacuum at 

 37 C. in the presence of caustic potash. The dried bacteria are 

 then ground in a mortar with the addition of a little dried sterilized 

 salt. Enough distilled water is then added to this powder to bring 

 the tonicity of the solution to normal (0.75 per cent). This emul- 

 sion is left until the following day, energetically centrifugalized, 

 and the supernatant fluid which is decanted is almost limpid. 



This fluid, so far as suspended materials are concerned, contains 

 sodium chloride in relatively large amount and bacterial substance 

 in proportionately small amount. And yet this fluid on intraperi- 

 toneal injection kills a guinea-pig in doses of 0.5 and even 0.25 of a 

 cubic centimeter, with the phenomena that we have just mentioned. 



Intravenous injection in a rabbit of 1 or 2 c.c. of endotoxin is 

 usually fatal, killing in less than 24 hours. At autopsy renal 

 hemorrhages, acute fatty degeneration of the liver and a marked 

 hemorrhagic condition of the suprarenal capsules are found. In 

 respect to this latter symptom, as well as in its tendency to pro- 



