450 INVESTIGATION BY CULTURE EXPERIMENTS 



tassium, in the standard forms (the nitrogen chiefly in sodium 

 nitrate). 



The treatment for these two plots differs only by the addition 

 of 13 pounds more nitrogen to plot n (plot n receives 50 pounds 

 of dried blood for wheat, and plot 20 only 25 pounds). The total 

 yields (except potatoes) are greater from plot n, and the total 

 value of the three crops is greater from plot n ($92.78) than 

 from plot 20 ($90.85); but the increase from plot 20 ($22.21) is 

 much greater than from plot n ($15.03), and the net profit from 

 20 ($11.08 or $19.94) is several times as great as from u ($1.95 

 or $6.91). In comparison with such plots as 14 and 27, it seems 

 evident that plot 20 gives results above normal; while it is like- 

 wise evident that plot u shows increases below normal, in com- 

 parison with such plots as 8 and 24. These opposite abnormalities 

 develop the striking discrepancy between n and 20. A study 

 of the yearly details shows that for the first five years (1895 

 to 1899) the increase from treatment was greater every year in 

 the wheat crops from plot u, the average difference being 3.6 

 bushels, whereas during the next five years (1900 to 1904) the in- 

 crease in wheat was greater every year on plot 20, the average 

 difference being 2.6 bushels. As an average of the four years, 

 1905 to 1908, the increase from plot 20 averaged 2.5 bushels more 

 wheat than from n, although in two of these years the treatment 

 gave about equal results on those plots. As an average of the first 

 six years (1896 to 1901) plot n produced 79 pounds less clover 

 than the unfertilized control plots, while on plot 20 the treatment 

 produced an average increase of 630 pounds. 



These results and discrepancies serve to emphasize' the uncer- 

 tainty of drawing correct conclusions from a single field experiment, 

 even when continued for several years. On the other hand, most 

 of the results from this potato-wheat-clover rotation are concord- 

 ant, and justify confidence. Indeed, there is marked agreement in 

 most cases where direct comparison is possible. Thus the increases 

 from like amounts of plant-food elements on plots 21, 23, 24, and 

 33 vary only from $18.18 to $18.78. 



In harmony with the results from all other sources, the use of 

 phosphorus on normal soils proves highly profitable, the increase 

 produced by phosphorus, both alone and in addition to other 



