ISOLATION OF B. TYPHOSUS FROM FECES 363 



obviously be dangerous to expose stools less rich in typhoid to 

 this marked absolute decrease in typhoid colonies. We next 

 tried shortening the time of shaking, and found that in no case 

 did the proportion of typhoid increase on shaking longer than 

 ten minutes, whereas the total number recovered frequently took 

 an alarming drop, and in some cases none was recovered after 

 thirty minutes. It was also determined at this time, that more 

 typhoid bacilH were viable after one than after two hours' 

 standing. The results with this modified technique— that is, 

 time of shaking reduced to ten minutes, and of standing to 

 one hour — may be summarized briefly as follows. 



These stools were quite rich in typhoid. Of those in which 

 bacilli were found both before and after shaking, the percentage 

 of typhoid to total flora was increased in three and decreased in 

 seven cases. In 3 of the 7, less than 10 per cent of the number of 

 colonies on the direct plates developed after shaking. In the 

 case of the 3 specimens showing relative increase, 2 of the 3 

 specimens showed loss of typhoid bacilli, in 1 case to 3 per cent 

 of the original number. The stool which gave positive results 

 only on direct plating had 75 per cent of typhoid (total colonies 

 900) and yielded 150 non-typhoid colonies after shaking. That 

 in which bacilli were found only after shaking had 6000 colonies 

 on two badly diffused direct plates, which were reduced on shak- 

 ing to 400, of which 20 were typhoid. In other words, the 

 tendency, even with the reduction of exposure, is to a rapid 

 reduction of the number of typhoid bacilli with which a relative 

 increase compared with the accompanying flora is not necessarily 

 associated. 



While these experiments were under way, we conducted a 

 number of tests with pure cultures of recently isolated strains 

 of typhoid, designed to develop a technique more favorable than 



