284 AGRICULTURE OF MAINE. 



best adapted to our conditions, and second, to originate new 

 varieties which, being bred in Maine, will be still better adapted 

 to our conditions. 



The work of breeding and testing new varieties is necessarily 

 slow. We have made some progress, however, and I will speak 

 of that in a few minutes. The work of testing existing varie- 

 ties goes more rapidly and we now have some results that 

 should be of value to the oat growler in the central and southern 

 part of the state. 



Before turning to the actual results of these variety tests I 

 would like to briefly outline the methods which we use in these 

 tests. In the first place it is necessary for us to test each year 

 a large number of different varieties. During the past season 

 we tested at Highmoor 53 distinct varieties or strams. Each 

 of these was tested in duplicate or quadruplicate. In all we 

 grew 150 plots in our variety tests this year. Each plot had to 

 be planted separately, harvested separately and threshed sepa- 

 rately. Obviously in work as extensive as this it is not possible 

 to grow large plots of each variety. Accordingly on the results 

 of our own experiments and those conducted at other places, 

 notably in England, we have adopted the plan of growing four 

 small plots of each variety. These plots are located in differ- 

 ent parts of the field. The average of these four plots is taken 

 as the yield of that variety for the given year. In all i-io of 

 an acre is devoted to each variety each year, but instead of 

 being in a single piece it is in four separate parcels. In this way 

 there is much less chance of one variety being seriously affected 

 by differences in the soil. 



These plots are planted with hand drills in rows six inches 

 apart. The seeding isi at the rate of two to three and one-half 

 bushels per acre. We have found that some of the larger grained 

 varieties must be seeded heavier than the small grained and this 

 accounts for the difference in the rate of seeding. The best 

 method of seeding in variety tests is not the same number of 

 pounds to a given acre, but the same number of kernels. We 

 have determined the average number of kernels in a given 

 weight of each variety and then plant such a weight of each as 

 will give approximately the same number of kernels. Thus a 

 large grained variety, like the Senator, must be seeded nearly 

 twice as heavy as a small grained one, like the Kherson. 



