276 BACTERIOLOGICAL EXAMINATION OF WATER. 



Proskauer and Capaldi's media, No. I. showed no growth, No. II. 

 was rendered acid. In gelatine plates, growth was slow, the 

 colonies were small, and resembled those of B. typhosus. Micro- 

 scopical examination showed a small highly motile bacillus, 

 decolorised by Gram. 



The tests showed that the bacillus which appeared in the 

 filtrate was identical with the B. typhosus planted out in the 

 sewage in the mantle. As the fittings were absolutely tight, 

 and the bacillus did not appear in the filtrate until the fifth day, 

 the experiment appeared to show that the sewage, just like the 

 broth, had supplied sufficient nutriment to enable the typhoid 

 bacillus to grow through the walls of the filtering candle. 



Experiment III. The same form of apparatus was employed 

 as in the two previous experiments. The mantle, however, was 

 filled with sterile barrack sewage, largely diluted with distilled 

 water. The dilute sewage was found to have the following 

 composition : Total solids, 7 parts ; chlorine, 0'3 part ; oxygen, 

 absorbed in four hours, 1-1 parts; free ammonia, (H248 part; 

 albuminoid ammonia,0'2016 part; nitrites, absent; nitrates (N() 3 ), 

 0*926 part; all expressed per 100,000. The dilute sewage was 

 inoculated with a sub-culture of the same typhoid bacillus and 

 filtration carried on daily ; the amount filtered was replaced by 

 sterile sewage, so that the filtering candle was never allowed to 

 become dry or exposed to the air. The broth in the Kitasato 

 flask remained perfectly clear for nine days, but on the morning 

 of the tenth day of filtration a slight turbidity was noticed. A 

 loopful of the turbid broth was removed through the arm and 

 planted out on agar. The growth which resulted was tested in 

 the same manner as before, and found to be a pure culture of 

 B. typhosus. The results of the agglutination and cultural tests 

 exactly corresponded to those of the B. typhosus introduced into 

 the mantle. In this experiment the typhoid bacillus was not 

 supplied with as much nutriment as in the two former experi- 

 ments, and it appeared to require nine days to grow through the 

 walls of the filtering candle. 



Experiment IV. The same form of apparatus was employed 

 as before, a perfectly new No. 12 candle being fixed in the 

 mantle. Surface water from a ditch was sterilised and placed in 

 the mantle. This water on analysis gave the following results, 



