54 MISSOURI AGR. EXP. STA. RESEARCH BULLETIN 49 



been used. Both of these are standard recommended varieties for the 

 region, though they differ decidedly in their characteristics. Both 

 have been used frequently as check varieties at the Missouri station in 

 past seasons. From the variability of the Kherson check plots the mean 

 yield of four replicate plots in this experiment would be considered 

 to have a probable error of 4.10 per cent ; from the Red Rustproof plots 

 the same determination would be given a probable error of 6.00 per 

 cent. A degree of precision for which we would assume four replica- 

 tions necessary, judging from the Kherson check, would require nine 

 replications according to the yields of the Red Rustproof check. 



The importance of choosing a check variety typical of the va- 

 rieties tested, if its variability is to be considered a criterion of the 

 variability of the field, is obvious. Whether it is possble to choose 

 a "typical variety" for the purpose, in the case of ordinary variety 

 tests, remains to be seen. 



ADJUSTMENT OF YIELDS BY MEANS OF CHECK PLOTS 



Adjustment of plot yields by the use of check plots has been a 

 common practice in field experiments during recent years. It is 

 recognized that no experiment field is perfectly uniform in produc- 

 tivity, and the attempt is made, by means of the check plot adjustment, 

 to compensate the varieties or treatments which chance to be located 

 on the less productive plots for the resulting loss in yield. The com- 

 mon method, in variety tests, is to distribute over the field, as fre- 

 quently as practicable, check plots planted to the same variety and 

 similarly handled in every way. The variation in yield among these 

 check plots is then considered a measure of the productivity of the 

 soil. By various methods, differing only in detail, the yields of the 

 test plots in parts of the field giving high check yields are reduced, and 

 those of test plots in parts giving low check yields are increased, in 

 proportion to the productivity of the soil, as indicated by the yields 

 of neighboring check plots. 



Previous Investigation. Several investigations of the effect of 

 such adjustment on the variability of replicate plots have been re- 

 ported. The majority of these have been conducted in connection with 

 experiments of the type discussed in the preceding section, in which 

 uniformly handled fields have been harvested in small sections. Cer- 

 tain of these sections, or plots, have been considered check plots, and 

 on the basis of their yields the yields of the remaining plots have been 



