228 BACTERIOLOGICAL AND ENZYME CHEMISTRY 



medium was determined at defined intervals. He afterwards 

 allowed the medium to drain and measured the amount of 

 drainage water, and finally dried the medium, and determined 

 the loss of moisture. He thus measured for each class of 

 medium (a) the amount of water passing through in a given 

 time, (b) the amount of water held in the larger interstices, 

 and (c) the amount of water retained in the pores. These 

 experiments showed generally that the time of percolation 

 through clean filter material varies, inversely as the rate of 

 application of the water, and directly as the amount of water 

 taking part in the water movement through the bed. 

 This latter obviously depends on the size of particles, and 

 the physical character of the medium. 



These results find expression in the following formula : 



~ 



where c is a constant, I the interstitial water per cubic 

 yard, B, the rate of sprinkling per square yard per hour, 

 and T the average time of sprinkling through three feet of 

 medium. 



Unpublished experiments by the author and Mr. T. W. 

 Lockett have shown that when a nitrifying solution, made up 

 after Winogradski's recipe, is allowed to drip on laboratory 

 filters, composed respectively of quartz particles about J inch 

 diameter, and of broken clinker of the same dimensions, nitri- 

 fication is established much more rapidly in the case of the 

 clinker medium than in the case of the quartz. 



De-nitrification. De-nitrification, as the name implies, 

 is the reverse of nitrification. De-nitrification changes are 

 concerned either with : 



(1) The reduction of nitrates to nitrites, or ammonia ; 

 (2) the reduction of nitrates and nitrites to oxides of nitrogen, 



