1912.1 PUBLIC DOCUMENT — No. 31. 21 



METHODS OF SELECTION FOR PLANT 

 IMPROVEMENT. 



BY J. K. SHAW. 



The fact that plants are unlike lies at the foundation of all 

 plant improvement. Sometimes it is first attempted to bring- 

 about increased unlikeness bj hybridization or otherwise, but 

 the great })art of the work is the choice of desirable plants and 

 by repeated selection securing and maintaining the desired 

 standard. Selection of desirable plants is generally followed 

 by .more or less positive results, but not uniformly so ; some- 

 times the results seem to be negative, or at any rate not as 

 satisfactory as might be desired. 



The plant is much the subject of its environment ; variation 

 in the nature and amount of the available food supply, and 

 the varying conditions of weather and climate, affect it strongly, 

 and selection of plants whose desirable qualities are thus brought 

 about must obviously be less effective than selection of differences 

 that arise from causes within the plant, and which persist from 

 generation to generation. In a collection of varying plant in- 

 dividuals some of the differences are environmental and are 

 inherited in small degrees if at all, while others are innate with 

 the plant and persist from generation to generation. The task 

 of the man who would improve his stock of beans or corn is 

 to distinguish between these two sorts of variation, and select 

 and hold to that which is not due to environment but is in- 

 herited by succeeding generations. 



Certain investigations that have been carried on by the de- 

 partment of horticulture during the past four years with garden 

 peas bear directly on this. This work is fully reported in Part 

 I. of this report (p. 82), to which the reader may refer for 

 detailed information. It is the purpose of this article to set 



