THE DISTRIBUTION OF MICRO-ORGANISMS l8l 



on the surface of the layer of fine sand. The filterbeds may 

 be several acres in extent, and in cold climates should be pro- 

 tected by arches of brick or stone. They require renewal occa- 

 sionally. This kind of filtration has come largely into use since 

 the cholera epidemic of 1892-93, and it appears to be very effective. 

 It is important to use storage basins in connection with sand 

 filtration. 



The results obtained by filtration depend greatly upon the 

 intelligence displayed in operation, and must be controlled by 

 frequent examinations of the water. 



Mechanical Filtration. This method of filtration is also 

 called the American system. It is more rapid than the preceding 

 method and does not require a large area for filter beds. Al- 

 though sand is required also, filtration is accomplished by a 

 jelly-like layer of aluminium hydroxide. This product is formed 

 by adding to the water a small quantity of aluminium sulphate 

 or of alum. The carbonates in the water decompose the aluminium 

 salt and produce aluminium hydroxide. It precipitates as a 

 white, flocculent deposit, entangling solid particles, including 

 bacteria, as coffee is cleared with white of egg. Only a 'trace 

 of aluminium should appear in the water. This method of filtra- 

 tion has not been tested so extensively as slow sand filtration, 

 but seems likely to prove efficient. With water poor in carbonates, 

 these may have to be added. 1 



Whipple and Longley 2 found that the efficacy of mechanical 

 filters with the addition of alum depends somewhat upon the 

 character of the alum. They find that the alum shall be shown 

 by analysis to contain 17 per cent of alumina (AljOj) soluble 

 in water, and of this amount at least 5 per cent shall be in excess 

 of the amount necessary theoretically to combine with the 

 sulphuric acid present. It shall not contain more than i per 

 cent of i soluble substances, and shall be free from extraneous 

 debris of all kinds. It must not contain more than 0.5 per 



1 See Fuller, Journal American Medical Association, Oct. 31, 1903. 



2 Jonrn. Infect. Diseases, Supplement No. 2, Feb., 1906, pp. 166171. 



