436 INVESTIGATION BY CULTURE EXPERIMENTS 



bushels on plot 106) to a point where perhaps the supply of avail- 

 able phosphorus again becomes the limiting factor. In other 

 seasons or in later years, these conditions may become reversed, 

 with nitrogen as the most limiting element, and phosphorus with 

 little or no effect except in addition to nitrogen. (Note the results 

 for 1907 and 1908, in Table 87.) 



We can conceive of conditions under which the supply of nitro- 

 gen naturally liberated from the soil, when supplemented by that 

 secured from the air by clover grown in the rotation, will meet the 

 needs of the crops grown for several years, during which the nitro- 

 gen does not become the limiting element to any marked degree, 

 and it must be plain that in such case the crop yields give little or 

 no information concerning the maintenance of nitrogen in the soil. 

 Thus, it is only after nitrogen becomes the limiting element, in 

 any given system, that the crop yields become an index as to the 

 possible permanency of the nitrogen supply. 



In soils that are markedly deficient in phosphorus, that element 

 may still remain the limiting element after the first small appli- 

 cation has been made, provided the increased supply of available 

 phosphorus is not sufficient to raise the crop yields to the point 

 where nitrogen, for example, becomes the limiting factor; and it is 

 easily conceivable that the increase produced by supplying po- 

 tassium in addition to phosphorus, in the Pennsylvania experi- 

 ments, was due, in part at least, to the power of potassium salts 

 to hold the phosphorus in available form. Even where heavy 

 applications of potassium were made, the sodium nitrate was more 

 effective than dried blood, and, if only sodium nitrate had been 

 added with phosphorus, the sodium would very probably have 

 produced nearly as marked results as were produced by potassium. 



There are too marked variations among duplicate plots on the 

 Pennsylvania field to justify fine distinctions, and even on more 

 uniform land there are many factors involved with different crops 

 and different seasons; but we dare not ignore the fact (Table 8iP) 

 that the average value of the crops from four acres receiving phos- 

 phorus-potassium treatment decreased from $88.32 to $80.27 

 during ten years, from 1891-1892 to 1901-1902, which are the 

 middle points of the two lo-year periods. Whether we consider 

 the values or the pounds of products, the apparent decrease is 



