14 BULLETIX 1092, U. S. DEPARTMENT OF AGRICULTURE. 



som Dutch variety, which is 6 inches shorter, but it also has a low 

 seed vitality and is inferior to the check in stem length. 



The third group, consisting of Selection Xos. 1904, 1919, 1923, 

 and 1931, appears inferior to the check but is analyzed in order to 

 make sure that no one of them is worthy of seed increase. Selec- 

 tion Xo. 1904 is the most promising one of this OToup in stem 

 weight : it is, however, inferior to the check in stem length, and its 

 resistance to wilt is low. Selection Xo. 1919 is very resistant to 

 wilt but its stems are coarse ; furthermore, when resistance to disease 

 is not considered it grades distinctly lower than the check. Selec- 

 tion Xo. 1923 is noticeably inferior to the check in length of stem and 

 is not deserving of general distribution ; it is, however, high in stem 

 weight in spite of its relatively short stems and may be suited to 

 areas where very tall growth is not desired to the exclusion of high 

 seed yield. Selection Xo. 1931 is too low in wilt resistance to war- 

 rant seed increase and distribution. 



It is concluded from a study of the score card and additional char- 

 acters which are not used in the score card that Selection Xo. 1905 is 

 the most desirable one for seed increase. Its strength of fiber is not 

 surpassed by any selection of which more than one year's record is 

 available. It is superior in resistance to wilt to all except two of the 

 eleven best selections, and these are undesirable because of coarse 

 stems and the extent of basal branching. Out of the five selections 

 that are its equal or superior in straw weight, two have a low seed 

 vitality, two are inferior in stem length and very low in wilt re- 

 sistance, and the remaining selection is more inclined to lodge than 

 the check. 



It is recognized that the percentages alloted in this score card are 

 moTe or less arbitrary and that, furthermore, they should be altered 

 to suit the convenience of the plant breeder. If, for instance, a 

 special attempt is made to select for resistance to lodging, it is 

 thought that a much higher percentage should be assigned to that 

 factor, so that it would be a determining one in deciding on the best 

 selection. 



The chief difficulty in the accurate working of the score card is 

 that the systems of planting as well as the places where the sowings 

 were made varied widely from year to year ; also, the records on all 

 the selections are not complete for the entire T-year period. Efforts 

 are made to overcome these difficulties by comparison with the check 

 which is planted each year in many duplicate plats. 



IMPROVEMENT BY CROSS-POLLINATION. 



An attempt has been made, beginning in 1918 and continuing up to 

 the present time, to combine the desirable qualities of several of the 

 best selections by cross-pollination. Third-generation progenies of 



