THE OUTLOOK FOR PLANT BREEDING. 99 



crosses. Again, in 1904, the best plants from the best progenies were 

 selected and these were planted in 1905 in a plat by themselves, 

 isolated from other cotton. This year the t}^e showed fair fixity 

 in all of its characters but the selection was again continued as 

 before, taking a number of the best plants from the best progenies. 

 Aside from these, selection was made of a considerable number of 

 tyi^ical plants to plant an increase plat the next year. 



In 1906, the special selections were planted again in an isolated 

 plat and fourteen acres were planted with the increase seed taken 

 from t\^Dical plants. The crop in this year showed greater uniform- 

 ity than in preceding years and was almost fixed in the character 

 of long lint. In the fall of 1906 selections were made in the breed- 

 ing patch of the best plants from the best progenies as before and 

 again a considerable quantity of seed was reserved for planting an 

 increase plat in 1907. The increase plat of 1907 contained over 

 twenty acres and the variety had become as stable in character as 

 any of the standard races. The lint averaged about one and three- 

 eighths inches in length, was strong and fine, and the plants were 

 large and productive. While the type of plant and boll was the 

 same as the original Russell, the lint was about one-quarter of an 

 inch longer and this greater length had become a fixed quality 

 or character of the new race. The race was thus markedly different 

 from the Russell and was named the Columbia, under which name 

 the new race was distributed. 



The original plant from which the Columbia came was thus, in a 

 sense, an accidental discovery. It seems probable that the careful 

 selection practiced for five generations in its improvement merely 

 had the effect of weeding out the results of crossing with inferior 

 plants, not of adding up any characters, for the lint of the first plant 

 found was as good as the lint of the finished product, after five years 

 of selection, during which time several thousand plants have been 

 handled each year. 



In the breeding of cereals and self-fertilized plants in general, 

 where the selected types would probably come true from the first, 

 if DeVries is correct, there would be no necessity of selecting the 

 best each year, but the possibilities of the improvement would have 

 been reached, with any race, when all of the t}^es had been sepa- 

 rated out and thoroughly tested as to their comparative values. 



