272 BACTERIOLOGICAL EXAMINATION OF WATER. 



also with water artificially infected with enormous numbers of 

 bacteria from cultures, and also with broth cultures teeming 

 with bacilli. 



(2) Water bacteria, however, begin to appear on the second 

 or third day of using, so that to obtain a continuously sterile 

 filtrate it is necessary to re-sterilise the filter daily. 



(3) A long series of experiments, in which London tap-water 

 was strongly infected with B. coli communis, showed that this 

 organism could not be detected in the filtered water over a 

 period of thirty-nine days. During this period the B. coli 

 communis could be detected on the outside of the filtering 

 candle as late as the thirty-ninth day. 



(4) Exhaustive experiments showed that with London tap- 

 water strongly infected artificially with the typhoid bacillus, not a 

 single typhoid bacillus was detected in the filtered water over n 

 period of twenty-six days, though it was detected in the 

 unfiltered water up to the sixth day. As is well known, typhoid 

 bacilli ultimately perish in water which contains the water 

 bacteria which are always present in natural water, but during 

 the whole time the typhoid bacilli were demonstrated to be alive 

 in the unfiltered water, they were never once detected in the 

 water which had passed through the filter. 



(5) London tap-water strongly infected artificially with the 

 cholera bacillus was found, after filtration through the filter, to 

 be absolutely sterile, and therefore entirely free from the cholera 

 organism. 



(6) The method of using filters with a combined suction and 

 force pump, causing as it does sudden and violent changes of 

 pressure on the outside of the filter, gives very bad results with 

 filtering candles, which produce an absolutely sterile filtrate when 

 used by a better method. The principle of the filters in which 

 atmospheric pressure forces the filtered water into an exhausted 

 receiver is free from the above objection, and when used by 

 this method absolutely sterile filtrates are obtained, even when 

 identically the same filtering candles are used which previously 

 passed large numbers of bacteria from the very first, when used 

 by the suction and force-pump method. 



(7) Seeing that the bacteria which first appear in the filtered 

 water are harmless water bacteria, and that such organisms as 



