PHENOMENA OF DRAINAGE WATERS 205 



soil (see p. 89). When no rain had occurred, the daily dis- 

 charge from the soil was small in quantity, and contained 

 very uniformly about 15 parts of nitrogen as nitrate per 

 million of water. This dry weather discharge would clearly 

 be derived from the lowest layer of saturated soil in the 

 drain- gauge. When however rain occurred in any consider- 

 able quantity, the strength of the drainage water immediately 

 fell, and only 10, 9, 8, or 7 parts of nitrogen as nitrate were 

 found in a million of water. The next day of dry weather 

 found however the strength of the drainage water nearly 

 re-established. The considerable weakening of the drainage 

 water during, or immediately after rain, was thus not due to 

 a washing out of the whole mass of soil, but to a temporary 

 dilution of the previous discharge with water of a different 

 composition. 



We will now give some further illustrations taken from 

 the facts observed when studying the drainage waters obtained 

 from the drain-pipes underlying the plots in Broadbalk wheat 

 field at Rothamsted (J. Roy. Agri. Soc. 1882, 15). These 

 plots are 8J yards wide,, and are of considerable length. 

 Under the whole length of each plot lies a drain pipe, about 

 2| feet below the surface. The lower ends of these drain- 

 pipes are uncovered, so that samples of the drainage waters 

 can be at any time obtained. The soil, as elsewhere at 

 Rothamsted, is a heavy loam. Below the subsoil of the field 

 lies the chalk. There is thus a good natural drainage, and the 

 water level in the subsoil probably never rises to the drain- 

 pipes, which are supplied with water from above. 



The waters collected from these pipes display remarkable 

 alterations in composition in different stages of the same 

 running. We shall find that these alterations are easily 



