THE ROTHAMSTED EXPERIMENTS 419 



amount applied as much or more than by the amount of crop har- 

 vested. Thus the organic matter from plot 7 is 1.54 times that 

 from plot 4-1; while the corresponding ratios are 1.62 for phos- 

 phorus and 1.66 for nitrogen, but 3.19 for potassium. Quite 

 similar results are secured by comparing plots 4-2 and 9, the ratios 

 beingi.56 for organic matter, 1.28 for nitrogen, 1.34 for phosphorus, 

 and 3.96 for potassium. 



These facts seem to harmonize with the suggestion previously 

 made that some of the apparent effect on the yield of hay, of the 

 potassium and other alkali salts, is associated with their power to 

 take phosphorus (in part from the surface, where applied in top 

 dressings) and deliver it to the root system of the plant. Another 

 influence of possible importance is the tendency of the alkali salts 

 to reduce or prevent soil acidity. Of course, potassium has some 

 value for its own sake under certain conditions, as clearly shown 

 in the previous and subsequent pages (see especially the mangel- 

 wurzel data). 



With the great differences that have developed in the character 

 of the herbage on the different plots, especially in different groups, 

 direct comparisons must involve several factors; and coincidence 

 or indirect correlation may easily be mistaken for direct causal 

 relationship. Thus, compared with plot 14, we would assume that 

 potassium must be the limiting element on plot 4-2 ; and possibly 

 such is the case, but reference to plots 7 and 10 show that other 

 factors are also involved. 



