PUEinCATION OF WATER FOR DRINKING PURPOSES 145 



separate determinations in each. case. The rate of fil- 

 tration was carefully regulated in each filter to a fall in 

 level of 26 to 28 centimetres per hour. 



The differences in the number of micro-organisms 

 found are so small that it is impossible to draw any 

 comparative conclusions from them ; but it is evident 

 that, at any rate with such pure water as was obtained 

 from the Lake of Zurich, it is possible for sand filters to 

 be kept working for from three to four years — which is 

 the period given in the report — and yet for them to 

 yield an exceedingly pure filtrate from a bacterial point 

 of view ; also that the simple renewal of the upper layers 

 of sand is sufficient to ensure the efficient working of 

 the filters. 



In connection with the influence of the thickness of 

 the layer of fine sand upon the filtrate, it is interesting 

 to note that in the course of removing the 70 cm. layer 

 of sand of one of the filter beds, which was yielding an 

 excellent filtrate, it was found that when only 25 cm. 

 was left the filtrate became unsatisfactory ; in conse- 

 quence of which it was decided not to reduce the layer 

 to less than 30 cm., but when this limit was reached in 

 the course of successive scraping, to commence the re- 

 newal of the sand. 



Some very interesting differences in the working of 

 covered and uncovered filter beds are also given ; thus 

 whilst the amount of water delivered by the former was 

 on the average 1,479,000 cubic metres per annum, the 

 latter yielded only 1,275,000 cubic metres, i.e, about 14 

 per cent, less, which must be attributed to their being 

 rendered less pervious by the more rapid formation of 

 slime through the growth of algae requiring daylight, 

 and to the effect of frost. Thus the surfaces of the 

 covered filters had on the average to be scraped seven 



L 



