PRINCIPLES OF HEREDITY 19 



remember that our new kinds need to be better than what 

 we now have. It is not enough that they be different. 



There are cases in which Mendel's law does not seem 

 to apply. Sometimes crosses do give blends, or inter- 

 mediates. Perhaps this is because we do not know what 

 unit characters are. 



19. Application of Mendel's Law. Since the results of 

 crossing give rise to such a miscellaneous array of forms, 

 only a very few of which are desirable, it is evidently not 

 a good practice for farmers to cross plants or animals 

 of different breeds. This is a very common practice of 

 American farmers, but certainly does not seem to be a 

 wise one. A man will get his herd of cattle graded up to 

 Shorthorn, then for some reason he changes to Hereford, 

 then to Angus, then perhaps back to Hereford. The re- 

 sult is a mongrel herd: It is much better to decide on a 

 breed and then keep breeding to pure-bred sires of that 

 breed. One will soon have a herd that is nearly pure-bred. 



Since we expect, not blends, but a recombination of char- 

 acters, we shall not expect to get a plant of intermediate 

 size by crossing a large one with a small one. Nor shall 

 we expect moderate-sized horses because one parent is 

 laige and one small. We are much more likely to get 

 an animal with bad proportions. We often see horses 

 with the body of a trotter carrying the feet of a draft- 

 horse and sometimes the head of a draft-horse. Many of 

 the ungainly horses that are seen everywhere are the 

 failures in attempts to get intermediates between distinct 

 types. (See Fig. 2.) 



If one wishes to produce an entirely new type or breed, 

 it is often desirable to cross unlike forms, in the hope of 



