QUANTITATIVE BACTERIOLOGICAL ANALYSIS. 39 



constructed that imperfectly filtered water cannot mix with tht,- 

 pure water. 



Reinsch and Kummel have devised an apparatus by which 

 samples of water for bacteriological examination can be taken 

 from filter-beds of old construction. The importance of being 

 able to examine each bed is well shown by the following results 

 of one of their investigations of filter-beds Nos. and 3 at the 

 Altona water-works ; No. 2 bed had been cleaned and a fresh 

 layer of sand 60 c.m. thick put on the bed : 



When examining the filtered water at least three gelatine 

 plates should be made and counts taken after forty-eight hours 

 incubation at 20 C. 



During the last two or three years the Massachusetts State 

 Board of Health has made experiments in order to determine 

 whether B. coli was present in the filtered water or not. "From 

 the studies of B. coli in filtered water it was thought probable 

 that a rate which would be entirely safe, judging only from the 

 determinations of the total number of bacteria present, might 

 really be too high to produce safe water, judging from the 

 B. coli determinations." The Board considers that the typhoid 

 bacillus cannot be present naturally in water without the 

 occurrence of large numbers of B. coli com munis, and regards 

 the absence of this bacillus from the filtered water as conclusive 

 evidence of the absence of the typhoid germ. But as 

 the Mem mack river, which receives the sewage from several 

 large cities, only contains fifty colon bacilli per cubic centimetre 

 at the intake of the Lawrence city filter, it is evident that the 

 typhoid bacilli contained in the river water must be very few in 

 number as compared with B. coli, even when the cities have 

 typhoid fever in an epidemic form. Consequently the volume 

 of water which must be declared free from B. coli in order to 



