174 MUTATION 



lows selection nearly or wholly ceases after four 

 or five generations, and if selection be abandoned 

 the race rapidly returns to its primitive condi- 

 tion. Such has been the experience of breeders 

 of maize, sugar beets, and other crops, and of 

 poultrymen who have sought to increase the egg 

 yield of fowl. Permanent improvement, wher- 

 ever made, has been eifected either by hybridiza- 

 tion with a wild form possessing the desired char- 

 acter or by preserving a fortunate sport — a 

 " Shakespeare," as Professor Hansen puts it. 

 Such a sport is a new center from which further 

 progress may start. Recognizing the futility of 

 selecting merely those individuals having the old 

 characters best developed, the most advanced 

 breeders (as at Svalof in Sweden), have system- 

 atized the search for single mutations in the 

 midst of extensive seed plats. This law of im- 

 provement holds for animals likewise. Four 

 years ago I started several series of experiments 

 to create, in poultry, new breeds by quantitative 

 selection. In one of these I sought to re-create 

 a uniform buff bird like the buffs that arose in 

 China two thousand years ago and are the par- 

 ents of all known uniformly red or buff breeds. 

 A bird with a red-and-black plumage coloration 

 of the Jungle fowl was crossed with a White 

 Leghorn. The hybrids were white with red on 

 the wings and breast. I then planned to breed 

 together the reddest of these birds and the red- 

 dest of their descendants until I should have 



