1 68 AGRICULTURAL BOTANY 



to breed. The planting of seed from a large squash is 

 not necessarily making a wise selection unless that 

 squash grew on a vine that bore more than one large 

 squash. The unit of selection should be the parent 

 plant, not one individual. 



Experiments have shown that there are two possible 

 ways to fix a variation : 



1. By propagating the plant by budding or graft- 

 ing (pages 159-165), instead of from the seed. This 

 method is practicable for trees, shrubs, fruits, potatoes, 

 and many flowering plants. 



2. By constant selection of seed toward the desired 

 type. This method is more or less practicable with the 

 grains and garden vegetables. For instance, if it is 

 desired to secure an especially early variety of corn, 

 we must save all seed from a single plant that early 

 matured large, strong ears, and plant them the following 

 spring. Not all the plants from these seeds will mature 

 even as early as the parent plant, but probably a few 

 will. The seed from these few should be saved and 

 planted the next spring. Continuing this process for 

 several seasons will in time fix a variety of early corn. 

 It must not be thought that the variety is thus per- 

 manently fixed, for the tendency is for such plants to 

 run out, that is, lose their distinctive characteristics. 



The only way to know whether improved varieties 

 will transmit their superior characteristics to the next 

 generation is to make tests. If, for instance, va- 

 rieties of corn have been produced which have full, 

 strong ears, a few should be planted from one ear of 

 each variety and the yields compared. In no other 

 way can it be determined whether selected individuals 

 will reproduce their desirable characteristics. 



