PURIFICATION OF WATER FOR DRINKING PURPOSES 167 



granular dust, through peat and through soil, the fol- 

 lowing may be taken as the percentage reductions 

 effected in the number of bacteria present : 



Very coarse sand, 5 feet in depth (grains averaging 

 about 0'06 inch in diameter), allowed from 26 to 40 

 per cent, of the bacteria to pass through. 



Finer sand, 5 feet in depth (grains averaging about 

 0*006 inch in diameter), 14 per cent. 



Very fine sand, 5 feet in depth (river silt), 5 per 

 cent. 



A mixture of coarse and fine sand and fine gravel, 

 3 feet 8 inches in depth, 5 per cent. 



Ditto, plus an upper layer of 10 inches of yellow, 

 sandy loam, and 6 inches brown soil, 5 per cent. 



Garden soil, 5 feet in depth. So very few organ- 

 isms were discoverable in the effluent that it is pre- 

 sumed that those found did not come through the filter, 

 but were due to post-filtration sources. 



Peat, 5 feet in depth. The latter remarks apply 

 equally to this filter. 



In the case of the two last-named filters, however, 

 the rate of filtration was so slow as to render them 

 entirely useless for practical filtration purposes. 



Purification of seivage by irrigation. In connection 

 with the above results it will be interesting to learn what 

 effect is produced on the bacterial contents of sewage 

 by its treatment in sewage farms. Miguel 1 has ex- 

 amined the sewage of Paris before and after it has been 

 led on to the land at Gennevilliers, and he finds that, 

 whereas the raw untreated sewage contains on an 

 average about 13,800,000 microbes per c.c., after 

 passing through the soil it contains on an average about 

 7,475. The effluent from this land is returned to the 



1 ' Manuel pratique d? Analyse bacteriologiquc des Eaux." 1 Miquel. 

 Paris, 1891, p. 13'J. 



