12 BULLETIN 33, U. S. DEPARTMENT OF AGRICULTURE. 



quotient is multiplied by the average yield of all check plats. The 

 two methods are more clearly shown in the following equations: 



FIRST METHOD. SECOND METHOD. 



a?=a+ ( 6 — y ) . x = a-r-jy x h, or ab^-y. 



Wherein. a= actual yield of variety. 



7)=average yield of all check plats. 

 #=coruputed yield of variety. 

 ?/=assumed yield of check. 



The check system as used at Dickinson has proved to be only 

 fairly satisfactory. It affords a means of gaining a better idea of 

 the relative yielding power of several varieties during certain seasons 

 than would have been the case had no system designed for that pur- 

 pose been used. In other seasons, however, the yields of the various 

 check plats were influenced by so many uncontrollable factors (mois- 

 ture, weeds, etc.) that to assume that any one of the other varieties 

 would be affected by the same cause and to a like degree would be 

 more of a theory than a determination. Usually the yields of the 

 check plats were quite uniform; however, they sometimes varied as 

 much as 50 to 100 per cent. A variety grown on a plat next to a 

 high -yielding check plat will have a computed yield less than its 

 actual yield. If next to a low-yielding check plat, the computed 

 yield of the variety will exceed its actual yield. When the yield of 

 a check plat is unusually high or low, the computed yield of varieties 

 grown on. adjacent plats are unusually and sometimes unreasonably 

 low or high. Such instances cause one to accept with caution results 

 based on computed yields alone. It is thought that actual yields of 

 a variety for a period of several years, especially when the variety 

 is shifted to several different places on the farm, are fairly depend- 

 able. For these reasons and because computed yields are not available 

 for all years, the conclusions in this bulletin are based principally 

 on the actual yields. The computed yields for those years in which 

 they were determined are given for the purpose of comparison. 



VARIETAL TESTS. 



The plat tests have included a large number of varieties and strains 

 of cereals imported from various parts of the world. Many of these 

 were introduced during the early years of the work and have since 

 been discarded. Those varieties which appeared to be best adapted 

 to the conditions of western North Dakota in regard to yield, growth 

 habits, utility, etc.. have been continued in the trials, which have 

 included 53 varieties and strains of spring wheat, 38 of spring oats, 

 16 of spring barleys, and a number of varieties of winter wheat and 

 rye. spring rye. and emmer. With some of these cereals the duration 

 of the trials is too short to be conclusive, while with others some very 

 satisfactory results have been obtained. 



