EXPERIMENT STATION REPORTS. 



305 



tion is based upon the fact that the plats are about twenty times as long as 

 the distance between checks, and therefore one is justified in assuming 

 that the soil fertility varies continuously between checks.^ The yield of 

 the check plats are determined in the same manner as those for the other 

 plats. The check-yields for the intervening plats are determined by 

 proportion of distance to the difference between the yields of adjacent 

 check plats. In the results that follow the figures in column P are the 

 actual 3'ields in pounds per plat for the varieties in question, and those 

 in the column C are the calculated yields of the check variety for the 

 same plats. In order to reduce this result to one figure, the yiejd is rep- 

 resented as a per cent of the yielding power of the check variety.' 

 It is desirable to run the entire series twice each year, and take an aver- 

 lOOP t^m,^ 



age between the ( ) coefficients of yield obtained for the duplicate 



C 

 plats. 



TABLE ]I. A PORTION OP THE .OAT^YIELDS 1911. 



P — Actual yields in pounds per plat. 



C — Calculated yield of check variety for each plat. 



Strain 

 No. 



61601 

 60401 

 60501 

 61201 

 61301 

 61601 

 61501 

 62704 

 62803 

 70503 

 61601 

 70806 

 76003 

 763001 

 76803 

 61601 



Variety. 



Check (Alexander).. 

 American Banner. . . 

 American Banner. . . 

 College Wonder. . . . 

 American Banner. . . 

 Check (Alexander).. 

 American Banner. . . 



Worthy 



Joanette 



Golden Fleece 



Check (Alexander) . . 



Clydesdale 



Improved American 



Joanette 



Progress 



Check (Alexander). . 



One Series. 



58.3 

 73.0 

 83.6 

 92.7 

 84.5 

 80.0 

 85.8 

 78.6 

 35.2 

 72.6 

 89.7 

 81.4 

 83.0 

 69.3 

 65.4 

 85.8 



C. 



100 P 



58.3 



62.64 



66.98 



71.32 



75.66 



80.00 



81.94 



83.88 



85.82 



87.76 



89.70 



88.92 



88.14 



87.37 



86.58 



85.80 



100.00 



116.54 



124.81 



129.98 



111.68 



100.00 



104.71 



93.71 



41.02 



82.73 



100.00 



91.54 



94.17 



79.33 



75.54 



100.00 



Duplicate Series. 



Ave. 



69.8 

 61.6 

 71.5 

 66.0 

 71.5 

 63.8 

 66.5 

 64.0 

 39.0 

 62.7 

 61.6 

 69.8 

 71.5 

 57.2 

 54.5 

 59.4 



C. 



69.8 

 68.6 

 67.4 

 66.2 

 65.0 

 63.8 

 63.36 

 62.92 

 62.48 

 62.04 

 61.60 

 61.16 

 60.72 

 60.28 

 59.84 

 159 . 40 



100 P 



C. 



100.00 



89.80 



106.08 



99.70 



110.00 



100.00 



104.96 



101.72 



62.42 



101.06 



100.00 



114.13 



117.75 



94.89 



91.08 



100 . 00 



100 P 



100 . 00 



103 . 44 



115.44 



114.84 



110.84 



100.00 



104.83 



97.71 



51.72 



91.89 



100.00 



102 . 83 



105.96 



87.11 



83.31 



100.00 



Table I presents four checks and the plats between them in the order 

 they were run in the 1911 variety series. It also gives the results of 

 the dupUcate portion of the series dealing with the same varieties. It is 

 easily observed that the values of C vary from one check to the next 

 according to distance and difference between the yield of adjacent checks 

 (They are straight line interpolations between checks). The results of 

 plant-row series are not expressable by this means, as their interpretation 

 is largely observational. The small plats of earlier years were nearly all 

 threshed with a flail. As losses are unavoidable by this means, the data 

 are not rehable plat results. The 1910 series was planted by the edge 

 of muck and certain areas lodged. 



» Rarely there is a reason to think that the zigzag fluctuations of the check yields does not 

 represent real variations in yield. In such cases a curve like y — a+bx+cx^ is used for 

 short distances. 



' This is known as the coefficient of yield and is determined by dividing the P by the 

 corresponding C, and multiplying by 100 to reduce to the form of per cent. If one wishes 

 these results In bushels per acre, it is only necessary to determine (the 100%). « reliable 

 tigura for the standard variety in bushela per acre. 



