160 BACTERIA IN WATER 



then allowed to fill, in order to get at what is really the im- 

 portant point, namely, the bacterial content of the water 

 entering the well. Again, if the sediment of the well has 

 been stirred up, a high value is obtained. Ordinary wells of 

 medium depth contain from 100 to 2000 per c.c. With regard 

 to rivers very varied results are obtained. Moorland streams 

 are usually fairly pure. In an ordinary river the numbers 

 present vary at different seasons of the year, whilst the pre- 

 vailing temperature, the presence or absence of decaying 

 vegetation, or of washings from land, and dilution with large 

 quantities of pure spring water, are other important features. 

 Thus the Franklands found the rivers Thames and Lea purest 

 in summer, and this they attributed to the fact that in this 

 season there is most spring water entering, and very little water 

 as washings off land. In the case of other rivers the bacteria 

 have been found to be fewest in winter. A great many circum- 

 stances must therefore be taken into account in dealing with 

 mere enumerations of water bacteria, and such enumerations 

 are only useful when they are taken simultaneously over a 

 stretch of river, with special reference to the sources of the 

 water entering the river. Thus it is usually found that im- 

 mediately below a sewage effluent the bacterial content rises, 

 though in a comparatively short distance the numbers may 

 markedly decrease, and it may be that the river as far as 

 numbers are concerned may appear to return to its previous 

 bacterial content. The numbers of bacteria present in rivers 

 vary so greatly that there is little use in quoting figures, most 

 information being obtainable by comparative enumerations before 

 and after a given event has occurred to a particular water. 

 Such a method is thus of great use in estimating the efficacy 

 of the filter-beds of a town water supply. These usually 

 remove from 95 to 98 per cent, of the bacteria present, and 

 a town supply as it issues from the filter-beds should not 

 contain more than 100 bacteria per c.c. Again, it is found 

 that the storage of water effects a very marked bacterial purifica- 

 tion. Thus Houston has shown in one series of observations 

 that while 93 per cent, of samples of raw river Lea water 

 contained b coli. in 1 c.c. or less, in the stored water 62 per 

 cent, of the samples showed no b. coli to be present in 100 c.c. 

 The highest counts of bacteria per c.c. are observed with sewage ; 

 for example, in the London sewage the numbers range from 

 six to twelve millions. 



Much more important than the mere enumeration of the 

 bacteria present in a water is the question whether these include 



