ACTION AND UTILITY OF FILTERS. 273 



B. coli communis and the typhoid bacillus do not accompany 

 them, even when they are present in large numbers in the un- 

 filtered water, it seems that in all cases in which a drinking- 

 water is not under suspicion of being contaminated, a semi- 

 weekly or weekly cleansing and sterilisation of the filtering 

 candle is all that is necessary as a precautionary measure 

 against water-borne disease. Nevertheless, in all cases of 

 contaminated water supplies, or supplies known to be infected 

 with disease organisms, a daily sterilisation of the filter- 

 ing candle is advisable, thus ensuring an absolutely sterile 

 nitrate. 



All the experiments hitherto described appear to prove con- 

 clusively that typhoid bacilli will not be able to grow through 

 either a Pasteur-Cham berland or Berkefeld filtering candle. 

 Unfortunately, my own work has led me to a different conclu- 

 sion in regard to the Berkefeld candles. The experiments I 

 made were as follows : 



Experiment I. A Berkefeld candle (No. 12 in the catalogue) 

 was placed in a mantle and firmly fixed in position by means of a 

 binding screw working on a thread on the metal pipe delivering 

 the filtered water. This metal pipe was connected, by means of 

 a short piece of stout india-rubber tubing previously fitted 

 with a pinch-cock, to a short glass tube which just passed 

 through an india-rubber cork fitting tightly into a Kitasato's 

 flask. The lateral arm of the flask, and the mouth of the 

 mantle surrounding the filtering candle, were tightly plugged 

 with sterile wool. The whole apparatus was then placed in the 

 autoclave and exposed to a temperature of 115 C. for half an 

 hour. The apparatus was allowed to cool, and the filtering 

 candle and mantle connected in the upright position to an 

 ordinary burette stand. The pinch-cock being placed in 

 position on the india-rubber tubing so as to separate the 

 mantle completely from the flask, all the joints of the appa- 

 ratus were thickly covered with melted paraffin. The mantle 

 was then filled with broth so as to cover the bougie completely, 

 but not touch the cotton \vool. Fifty cc. of broth were placed 

 in the flask and, the bung being replaced, the joint was made 

 tight with melted paraffin. These manipulations were done 

 with every precaution, and carried out as quickly as possible, 



