CROSSING. 63 



very good example. Nearly every cross between animals and 

 plants, which are not too closely related produces new charact- 

 ers as its result. 



' OOOOP'F, 



Cf 



Fig. 8. 

 Inheritance of albinism in a family of Malayan field-rats. 



We do not think it probable, that new breeds of animals or 

 plants, other than those which are habitually self- fertilized, are 

 directly derived from hybrid individuals. Even in the history 

 of the tame species of fowls, not many instances are known, 

 in which a new breed was deliberately produced by crossing, 

 out of the variable off- spring of hybrid birds. The common way 

 in which new breeds of fowls are made, is by breeding a new 

 character into an old species. Species with a colour, new for the 

 group, are often produced, by breeding an animal of the desired 

 colour with good typical representatives of the species, and by 

 continuing to breed the hybrid back to the old species until 

 the desired result is obtained. It is especially easy to introduce 

 a gene, not heretofore present, as the desired character which 

 in this species greatly depends upon the presence of the gene, 

 is not lost sight of. In our size-inheritance work with mice, we 

 find it extremely easy to retain any one gene present in one 

 species during a four or five times repeated crossing back, 

 "grading up" to a second species. By breeding back to an 

 albino strain, without more selection than retaining any pig- 



