152 DOMESTICATED ANIMALS AND PLANTS 



Such radical departures frorn type were formerly recognized 

 and popularly designated as "sports," as if nature in some 

 sudden antic disposition, at play in her workshop, were disre- 

 garding all ordinary laws of procedure. 



The modern name of mutants is better, and while these sud- 

 den departures are often independent of crossing, it is signifi- 

 cant that they frequently breed true, showing that the changes 

 involved are sufficiently profound to affect the germ plasm. 



The selection and isolation of desirable mutants, therefore, 

 constitutes a third method of improvement of animals and plants, 

 the one most practiced by Luther Burbank. 



Origin of new and improved strains. Three methods of im- 

 provement are therefore open to the breeder: (i) selection in 

 imitation of nature ; (2) crossing, with the understanding that 

 new strains may also be shaped up by selection ; (3) mutation, 

 the fortunate mutants being seized upon and made the most of 

 as a free gift of nature to the breeder's hand. 



Doubtless all these means of changes in species are in opera- 

 tion everywhere in nature. Darwin expounded the first and 

 De Vries the last, and a multitude of evolutionary literature 

 exists. The student who is desirous of pursuing the general 

 question of origin of species in nature will find the subject 

 briefly sketched in Chapters XVII-XXI of Part II, with some 

 standard references. 



Summary. When distinct races are crossed hybrids are produced be- 

 tween all the characters involved. Some of these characters will blend, 

 and the result will be a new combination which will thereafter breed true 

 as regards all such blended characters. 



But other characters will not blend, remaining distinct, in which case the 

 gametes will continue to produce not one new and blended character, but 

 both old characters in their original purity. Under the law of chance one 

 fourth of the offspring would possess the character of the one parent in its 

 purity, one fourth that of the other, and half would remain hybrid. 



Inasmuch as some characters are naturally dominant and others recessive, 

 the recessive individuals can be detected only where the recessive stands 



