Plant Breeding. 271 



guarded, and when the art of propagation became 

 known, these plants were most multiplied. In each 

 successive generation, the most desirable individual 

 plants of each species were protected and multiplied, 

 or at least were permitted to perpetuate themselves. 

 Since the offspring tends to resemble the parent (18), 

 the persistent propagation from the best has resulted 

 in more or less marked improvement. Chance cross- 

 ings have aided the process (445). These facts fur- 

 nish hints for the further improvement of plants. 



434. The Variability of plants Renders their Im- 

 provement Possible. ~* In a species of which the indi- 

 vidual plants are all practically alike, as in many wild 

 plants, we can do little in the way of plant breeding, 

 except to give treatment that promotes variability. In 

 a species in which the individuals manifest different 

 qualities, however, we may hope to secure improvement 

 by using the more desirable plants as parents from 

 which to secure still further variability. 



435. Variations are Not Always Permanent. If we 

 find a chance seedling of the wild blackberry, for ex- 

 ample, that has remarkably fine fruit, the plants 

 grown from seeds of this fruit are not always equal in 

 quality to the parent. The tendency, in such cases, is 

 for the seedling plants to revert or go back to the ordi- 

 nary type of the species, and the more marked the 

 variation, the stronger is the tendency to reversion. 



436. How to Fix Desirable Variations. A fixed 

 variation, i. e., a variation of which the progeny re- 

 sembles the parent in all important characters, be- 



