196 GENERAL BOTANY 



variations is followed in the improvement of other kinds of fruit, 

 forage, and food-producing plants. The following extract from 

 " Plant Breeding," by Professor Bailey, illustrates some of the 

 detail and care necessary in such a selecting process when applied 

 to the cultivation of beans. 



He [the breeder] starts with one plant. The next year he may 

 have only two. If he has ten or twenty good ones, then the task is 

 easy, for the variety has elements of permanence, that is, of heredi- 

 tability, in it. As soon as seeds can be secured in considerable 

 amount from a strain of beans, selected as indicated above, the 

 grower can plant a large plot and obtain seeds for sale or for dis- 

 tribution to other breeders. He must exercise judgment and skill 

 each year, however, in selecting seed, even from such a carefully 

 improved race, if he is to prevent its running out or reverting to 

 the original type. This is due to the fact that his fields and experi- 

 mental plots will always contain reversions, or " rogues," from which 

 seed must not be taken if his improved strain is to be kept true to 

 type. In the common dwarf, or bush, beans of the gardens, for 

 instance, there is always a tendency to revert to the ancestral climb- 

 ing variety. These rogues with a tendency to climb must, there- 

 fore, be eliminated if the bush habit is to be perpetuated. 



Mr. Palmer's dwarf lima originated in 1883, when his entire crop 

 of large white (pole) limas was destroyed by cutworms. He went 

 over his field to remove the poles before fitting the land for other 

 uses, but he found one little plant, about ten inches high, which had 

 been cut off about an inch above the ground, but which had rerooted. 

 It bore three pods, each containing one seed. These three seeds 

 were planted in 1884, and two of the plants were dwarf like the 

 parent. By discarding all plants which had a tendency to climb, 

 in succeeding crops, the Burpee bush lima, as we now have it, 

 was developed. 



Difficulties. Two main difficulties are experienced by breeders 

 in their attempt to improve plants by the selection of variations. 

 One of these difficulties has already been referred to, namely, 

 the tendency of varieties and races thus produced to revert to 

 the condition of ancestors of the less desirable type, a phenomenon 

 expressed by the term running out. A second point of impor- 

 tance is that races built up by the continued selection and 



