300 ACTIONS OF ARTIFICIAL MANURES 



must be some substance in the soils from the nitrated plots 

 which had brought them into a deflocculated condition, and 

 this substance could not well be anything else than a trace 

 of alkali. On testing, the soils from the nitrated plots 

 were found to be slightly alkaline, probably with carbonate 

 of soda, whereupon a quantity of soil from the nitrate plot 

 on the grass land was washed with hot water to see how 

 much alkali could be extracted from it, this particular plot being 

 selected because the soil contained no carbonate of lime, which 

 itself might give rise to a soluble alkali. Table C. shows 

 the quantities of carbonate of soda that were found in the 

 successive 9-inch layers down to a depth of 3 feet, the results 

 being calculated as Ib. of carbonate of soda per acre. 



TABLE C. Carbonate of Soda (Ib. per acre) in Soil of Plot 14, 

 Park Grass Field, Eothamsted. 



It will be seen that the total amounts to no less than 175 Ib. 

 of carbonate of soda, which is the chemical equivalent of 280 Ib. 

 of nitrate of soda, or about one-half of the yearly application 

 (550 Ib. per acre) of nitrate of soda to this plot. We cannot 

 base our calculations on more than the year's application, because 

 neither nitrate nor carbonate of soda are in the least retained 

 by the soil, and both must wash out pretty completely during 

 a wet winter. The problem then was thus far cleared it had 

 been shown that the soils which receive nitrate of soda after- 

 wards contain carbonate of soda equivalent to as much as 

 one-half of the nitrate applied, and this carbonate of soda was 

 in itself enough to account for the bad texture of the soils, a 

 bad texture which is due not to any special defect of com- 

 position but to the deflocculation of the clay particles in the 

 soil. The next question was to account for the formation of 

 the carbonate of soda, and here certain well-established facts 



