EXPERIMENT STATION REPORTS. 



309 



A general view of the 1911 variety series is shown in figure four. It 

 will be noticed that the plats were full drill widths (eleven hoes), at that 

 time and that the plats were much shorter than now. That shape of plat 

 was used during the years of 1909 to 1913. The duplicate series of 1911 

 were end to end and ran in reverse order. During recent years, the du- 

 plicate series have been parallel, i. e., a continuation of the same series. 

 This permits the plats to be longer and narrower, bringing the check plats 

 nearer together and making the determination of the check yields for the 

 intervening plats more reliable. The best results are obtained when the 

 length of the plats are about twenty times the distance between the cen- 

 ters of the check plats. The accuracy of the results are indicated by the 

 nearness of the coefficients of yield representing the duplicate plats. 

 They should usiinlly be within 5% or more rarely within 10% of each 



Fig. 4. Up to 1913, the plats were full drill width. Figure 4 shows a general view 

 of the 1911 oat series. There were two series end to end that year, the plats only being about 

 300 feet long. That left the length of the plats about 14 times the distance between the 

 centers of the check plats. This relation should be at least twenty times. 



other. A difference of 15% or 20% means a mistake that is not due to 

 the relationship of the plats if the plats are arranged as above. A patch 

 of weeds, or imperfect planting in one of the duplicates may cause a dif- 

 ference of 30%. Shattering in spots if over-ripe, or lodging in small 

 localities may create a large mistake. These mistakes also arise at har- 

 vesting and during threshing if more grain is lost from one plat than 

 is lost from another. If the loss is equal from all the plats (checks and 

 all), the results (coefficients of yield) are disturbed very little if at all. 



In calculating, four place figures have usually been made use of, espec- 

 ially in the yield for the check variety on the intervening plats. These 

 four place figures have been used in the calculation of the per cent or 

 lOOP 



. Also averages are obtained with four place coefficients of yield. 



C 

 To save space the results are given in the tables as three place figures. 



I 



