1905.] 



On the Retention of Bases by the Soil. 



19 



Annual loss of calcium 



Broadbalk (Plot 7 - Plot 3) -=-2 .. 



(Plot 10 -Plot 3)-r-2.. 



(Plot 8-Plot3)--3.. 

 Hoos (Plot 4a -Plot 4o)-- 1 



Mean 



117 



This mean value for the loss caused by 200 lbs. of ammonium salts is not 

 very far from the 161 lbs. estimated above as necessary to convert them into 

 ammonium carbonate, but neither the mean nor any of the individual 

 analyses support the view that a second molecule of calcium carbonate is 

 removed from the land by the nitrification of the ammonium carbonate. 

 However, this nitrification must take place ; indeed, there is every evidence 

 that it takes place so rapidly and thoroughly in the Kothamsted soil that no 

 ammonium salts are carried forward un-nitrified from one season to the next. 

 The analyses of the drainage waters (Table X), while they show a progressive 

 increase in the amount of lime for each addition of ammonium salts in the 

 manure, do not permit any estimate to be formed of the rate of removal, so 

 much is the extent of the percolation, as seen in the relative frequency with 

 which the drains run, affected by the size of the crop, which becomes large on 

 the heavily-manured plots. It remains, therefore, to be explained why the 

 loss of the soil should be at the rate of one rather than of two molecules of 

 calcium carbonate for every two molecules of combined ammonia applied in 

 the manure. 



The plots receiving sodium nitrate in place of ammonium salts show not 

 only no special loss of calcium carbonate due to the nitrogeneous manure, but 

 a distinctly diminished loss as compared with the unmanured plot. On 

 Broadbalk the nitrate plot loses at the rate of 564 lbs. per acre against 

 800 lbs. on the unmanured plot ; on Hoos the loss is 465 lbs., against 675 lbs. 

 on the plot receiving the same minerals but no nitrate. 



That the sodium nitrate exercises some conserving influence on the calcium 

 carbonate in the soil is also apparent from a consideration of the analyses of 

 the subsoil. On Broadbalk, where the subsoil from 10 to 18 inches contains 

 in most cases 0*1 or 0'15 per cent, of calcium carbonate, the same layer 

 beneath the plot receiving sodium nitrate shows 0*31, 0*26, - 48, and 

 - 14 per cent, at the various dates, while the third depth of this plot is also 

 richer than on the corresponding plots. 



Further evidence may be derived from the composition of the drainage 

 waters (Table X) ; the water from the plot receiving mineral manures only 



c 2 



