BACTERIUM TYPHI. 239 



lets or exceedingly finely granular, contrasting with the 

 large, markedly granular, brown-colored colonies of the 

 Bact. coli. 



The method is said to give ver}"- good results, and usu- 

 ally allows of the isolation of the typhoid bacterium from 

 stools, and the results are said to be most perfectly in har- 

 mony with Pfeiiier's typhoid reaction (see below). Com- 

 pare also Jemma (Miinch. med. Woch., 1897, No. 33) 

 and Sterling (C. B. xxii, 334). 



Special Differential Diagnosis of the Bact. typhi, Es= 

 pecially from the Bact. coli. 



The following peculiarities must all be demonstrated : 



1. Rods, short to thread forms; active motility; abun- 

 dant, long, peritrichous flagella; not stained by Gram's 

 method. 



2. White film upon gelatin which is not liquefied. 



3. No formation of gas from grape- or milk-sugar in a 

 shake culture. 



4. Uniform cloudiness of sugar bouillon in fermentation 

 tubes without formation of gas. No formation of acid 

 from milk-sugar, abundant from grape-sugar. 



5. No coagulation of milk. 



6. Indol not produced in peptone solution. 



7. Finally, Losener places value upon the demonstration 

 by means of cultures in Petruschky's litmus whey (at 37°) 

 that the questionable typhoid bacterium in about forty- 

 eight hours does not produce more than 3.0 c. c. of deci- 

 normal acid from 10 c. c. of milk, while the coli bacteria 

 form more than 8 c. c. i 



8. Marked agglutination by specific serum (see below). 



9. Of less value in the diagnosis are: (1) The microscopic appear- 

 ance of the gelatin plate, as it may be almost identical with the Bact. 



' Upon all these points a very satisfactory nniformity has been 

 reached. To be sure, the uniformity depends in part upon an agree- 

 ment, which is, that all those bacteria which do not present these pecu- 

 liarities of the typical typhoid culture are simply declared to be differ- 

 ent from typhoid, under the assumption that the typhoid bacterium 

 does not vary. How little probability this assumption possesses in the 

 face of the enormous variability of the closely related Bact. coli, re- 

 quires no discussion. 



