12 BULLETIN 450, U. S, DEPARTMENT OF AGEICULTUEE. 



The Haynes lias the highest average yield of flour, volume of loaf, 

 and texture of loaf. The Rysting-Glyndon has the highest average 

 crude-protein content and color of loaf. The Kubanka shows the 

 smallest loss in milling. 



The data given show that the Haynes is superior to the Ghirka in 

 all of the six characters studied. The Rysting-Glyndon is superior 

 in all characters except loaf volume, and the Kubanka is superior m 

 all characters except volume and color of loaf. 



IMPROVEMENT BY SELECTION. 



The high-yielding and drought-resistant power of the Ghirka wheat 

 was early recognized at the Dickinson substation. A further improve- 

 ment in the yield and quality of the variety was sought through 

 isolatmg a large number of pure lines and comparing them with the 



Fig. 5. — Pure-line selections of Ghirka Spring wheat growing in eightieth-acre plats at the Dickinson 



substation in 1913. 



original. This work was started in 1909, when about 300 heads 

 were selected from a plat of the Ghirka wheat.* The kernels from 

 these were sown in head rows in 1910. Yield data were obtained, 

 and all the best, a large number of the medium, and some of the 

 poorest were selected for sowing in 1911. Of these selections, 104 

 were sovm in 17-foot rows for comparative 3'ields, with the original 

 Ghirka mass variety used as a check every third row. Eighty of 

 the selections were sown also in rows of varying length to increase 



1 The experiments to improve the Ghirka Spring variety by pure-line selection were suggested by Supt. 

 1.. R. AValdron, of the Dickinson substation, who has maintained his interest in the -w^ork throughout 

 the entire period. The original selections were made by Mr. Charles H. Clark, who was in charge of 

 the cereal work at the Dickinson substation inl909and 1910. From 1911 to July 25, 1914, the writer was in 

 direct charge of the cooperative cereal experiments at Dickinson. On the latter date he was succeeded 

 by Mr. Ralph W. Smith, who continues in charge of the work as Scientific Assistant in Cereal Investiga- 

 tions. 



