THE ROTHAMSTED EXPERIMENTS 415 



age occurs when the land is not covered with vegetation. When- 

 ever roily water leaves a field, some soil goes with it; and the loss 

 of a tenth of an inch in three or four years is not improbable, even 

 for nearly level land, if annually cultivated, especially if torrential 

 rains sometimes occur (see record of Barn field) 



Whether one assumes 10 inches or 13^ inches of drainage, there 

 is some degree of correlation between the computed calcium 

 carbonate equivalent to the calcium found in the drainage water, as 

 shown in Table 74, and the loss of calcium carbonate from the sur- 

 face soil of Broadbalk field, as recorded in Table 27. While there 

 are marked discrepancies, both methods agree that, as an average, 

 more calcium is removed from the plots receiving ammonium 

 salts. 



Analyses made of surface soil from the barley plots on Hoos 

 field in 1889 show in 2 million pounds of soil 960 pounds of phos- 

 phorus as an average in the 8 plots receiving no phosphorus, 1560 

 pounds as an average in the 8 plots receiving acid phosphate with- 

 out rape cake, 1900 pounds as an average in the 2 plots receiving 

 acid phosphate and rape cake, and 1540 pounds in the farm manure 

 plot (7-2). 



Table 75 shows the nitrogen content of the surface 9 inches of 

 the different plots on the Agdell rotation field. 



From the data thus far reported, the nitrogen content of the soil 

 on Agdell field appears to be decreasing about 10 pounds a year, 

 except on the legume plots which receive rape cake and ammonium 

 salts, where an increase is shown on the " fed " plot amounting to 

 212 pounds in 16 years. While the individual variations are great, 

 the results indicate a slightly larger loss of nitrogen in the legume 

 rotation than with fallow, but where nitrogen is applied, the op- 

 posite is shown. 



The factors of erosion and deposition and of difficulty in securing 

 samples (by the method used) which fairly represent the average 

 of the plot are sufficient to account for any of the changes indi- 

 cated by these analytical data; and it may be stated that the to- 

 pography of Agdell field suggests the possible influence of such 

 factors. On the other hand, the indicated gain of 180 pounds of 

 nitrogen per acre during seven years with the legume rotation on 

 the unfertilized land, with all crops removed, has actually been 



