72G A MANUAL OF BACTERIOLOGY 



The Berkefeld, while more rapid in action than the 

 other two, after being in use for a few days may allow 

 some organisms to appear in the filtrate. This, perhaps, 

 is due to a growth of organisms through the pores 

 of the filter-candle rather than to a direct passage. 

 Lunt l found that while the ordinary water bacteria, 

 such as the B. fluorescens liquefaciens, appeared in the 

 filtrate from a Berkefeld filter within a few days of the 

 infection of the sample, the typhoid bacillus and the 

 comma bacillus similarly introduced had not passed 

 through the filter four or five weeks after infection. 



Horrocks, 2 however, found that when sterile water is 

 inoculated with typhoid bacilli and run daily through a 

 Berkefeld filter, the bacilli appear in the filtrate in one or 

 two weeks, whereas this is not the case with the Pasteur- 

 Chamberland. The writer has made some similar experi- 

 ments, which partially, but not entirely, support Hor- 

 rocks' s conclusions. Much evidently depends upon the 

 chemical composition of the water. 



All porcelain filters should be cleaned weekly by well 

 scrubbing with a nail-brush and boiling in water contain- 

 ing some sodium carbonate. 



The Bacteriological Examination of Water-Filters 



In order to ascertain whether organisms pass through a filter, 

 it should be sterilised in the steam steriliser, and sterile water 

 infected with organisms of known species (B. prodigiosus, 7>. 

 violaceus, and M. agilis are very suitable) should be passed through 

 it for twenty-four hours. This water and the filter should during 

 this period of the examination be maintained, if conveniently 

 possible, at a temperature below 5 C. This will almost invariably 

 prevent any growth or multiplication of the organisms. Samples 

 should be taken immediately after the filtration has begun, and 

 . 



1 Trans. Brit. Imt. of Prev. Med., vol. i, 1897. 



2 Brit. Med. Journ., 1901, vol. i, p. 1471. 



