PURCHASE: AND USE OF FERTILIZERS 



327 



parison shows the differences between these, and combination and 

 comparison shows the differences of larger areas. For example, 

 Morgan 1 selected a strip of land 112^2 feet wide apparently uni- 

 form in texture, etc., which was planted in wheat first and then 

 corn. It was measured off in strips 15x112^ feet, (about 1/25 

 acre) and the wheat or corn harvested separately from 63 plots 

 which should thus be all alike. With the average yield at 100, 

 the wheat crop varied from 65.0 to 130, and the corn crop from 



169.3 to 42.3. 



The plot yielding the lowest with wheat gave an average yield 

 with corn, the highest with wheat 88.5 with corn. The plot lowest 

 with corn was 1 13.7 with wheat, and the plot highest with corn was 



100.4 with wheat. Thus the differences were not in the same 

 direction with the two successive crops. Assuming that the 

 theoretical yield depended upon the distance from the check plot, 

 the following average errors were found : 



Assuming different treatments for the plots, the average and 

 maximum error was found to be as follows : 



Proc. Am. Soc. Agri., 1909. See also Lyon, ibid., 1910, p. 35-38. 



