ACTION AND UTILITY OF FILTERS. 219 



Experiment XI. In this experiment the bougie used in 

 Experiment VII. was sterilised in an autoclave and then fitted up 

 in a sterile apparatus. Sterile reservoir water was filtered through 

 into the flask for eleven days without the slightest trace of 

 turbidity appearing in the broth. 



Experiments on the same lines were then made with Pasteur- 

 Cham berland candles, fitted into glass mantles by means of 

 india-rubber bungs. The delivery pipe of the candle in each 

 case was connected by a short piece of india-rubber tubing, fitted 

 with a pinch-cock, to a Kitasato flask containing sterile broth. 

 As in the previous experiments, all the joints were made tight 

 with melted paraffin. In the first trial, the mantle was filled 

 with sterile broth, which was then inoculated with a twenty-four 

 hours culture of B. typhosus. Next day the broth in the mantle 

 was extremely turbid ; filtration was then commenced, 10 c.c. of 

 broth being filtered into the flask every day and 10 c.c. of sterile 

 broth being put into the mantle. After filtration for three 

 weeks, the broth in the flask was found quite sterile. One loop- 

 ful of the broth from the container, however, contained a large 

 number of typhoid bacilli. Trials were then made with the 

 other candles ; sterile barrack sewage, sterile sewage effluents, 

 polluted ditch water and reservoir water, repeatedly inoculated 

 with large doses of B. typhosus, were filtered for three weeks, 

 but no traces of B. typhosus could be detected in the broth 

 placed in the Kitasato flasks. The failure to pass through the 

 filtering candles was not due to the absence of B. typhosus 

 from the fluid in the mantles, as in each experiment one loopful 

 of the fluid in the mantle at the end of the trial was found to 

 contain large numbers of living, vigorous, typhoid bacilli. 



The following conclusions appear to be ju^tilied . 



(1) The B. typhosus is not able to grow through the walls of a 

 Pasteur-Cham berland candle, and, if proper care be taken to pre- 

 vent the direct passage of organisms through flaws in the material 

 and imperfections in the fittings, the Pasteur-Chamberland filter 

 ought to give complete protection from water-borne disease. 



(2) Typhoid bacilli can grow through the walls of Berkefeld 

 candles, the time required for the passage being largely dependent 

 on the nutriment supplied to the organisms by the filtering fluid. 

 The large size of the lacunar spaces in the Berkefeld candle, 



