45 



that the figures correctly r ^present, in degree, the average difference 

 in the composition of the first 9 inches of the two plots ; for, calcu- 

 lated per acre, the excess of nitrogen in the surface soil of the clover 

 plot would represent an accumulation equal to about twice as much 

 as was removed in the three cuttings of the clover, notwithstanding all 

 visible vegetable debris was removed before the soils were submitted 

 to analysis ;* nor have the subsequent crops benefited as much as might 

 have been expected from such an amount of accumulation. On the 

 othtir hand, samples taken in 1877 still show a higher percentage of 

 nitrogen in the surface soil of the clover than of the barley land. 



It is, at any rate, obvious that the surface soil of the clover ground 

 has gained nitrogen, either from above or from below — from the 

 atmosphere or from the subsoil. And, so far as the determinations of 

 nitrogen in the subsoils go, the indication is that, if from below, it is 

 at least mainly from a lower depth than 27 inches. 



It is freely admitted that, in the facts of this experiment as they 

 stand, there is no evidence as to the source of the large amount of 

 nitrogen of the clover crop, and of the increased amount of it 

 in the surface soil. In the absence of such evidence, it is natural 

 enough to assume that the atmosphere has been the source. But 

 whilst there is absolutely nothing in favour of this view excepting 

 the fact that an explanation is needed, and that if that source were 

 established the difficulty would be solved, there is, to say the least, 

 much more evidence in favour of the supposition that the subsoil has 

 been the source of at any rate much of the nitrogen. 



The Soils of the Melilotus leucantha and White Clover Plots. 



Reference has already been made to the enormous growth of 

 Melilotus leucantha, and the enormous amount of nitrogen it yielded, 

 for several years in succession, on the land where no nitrogen had 

 been applied for more than thirty years, and where red clover had so 

 frequently failed (p. 12). The crop of 1882, the fifth in succession, 

 was the highest, and the yield of nitrogen in it was not far short of 

 150 pounds per acre ; whilst, under exactly similar conditions, ordi- 

 nary red and white clover gave very small produce. Accordingly, as 

 soon as the crops were removed, samples of soil were taken from one 

 of the melilotus plots, and from the corresponding white clover plot. 

 Samples were taken from two places on each plot, and in each case to 



* This was more completely done in the case of the four 12x12x9 inch samples, 

 than in that of the six 6 x 6 x 9 inch ones, and the latter are seen to give slightly 

 higher percentages of nitrogen. 



