GENERAL CONSIDERATIONS AND CONCLUSIONS 441 



operations of breeding result in "new creations" in the sense of bringing 

 new characters into existence outside of those already attained in the 

 course of evolution. Moreover, it would be very difficult to maintain a 

 distinction between conservation and construction in modern breeding. 

 A scientifically planned system of breeding improved crop plants must 

 be founded upon suitable methods of testing the species, varieties and 

 biotypes available in order to discover the most promising material 

 for the constructive work of the hybridist. 



Uneconomical methods must be eliminated. There are sources of 

 enormous waste in present day plant breeding work. An example is 

 the useless attempt at improving pure lines through continual selection; 

 another is the assumption that a single pure line selection represents all 

 there is in a variety, a serious mistake since most commercial varieties 

 of self-fertilized plants consist of a mixture of pure lines. The frequency 

 of mutations in all economic plants has a direct bearing upon these ques- 

 tions of breeding practice; hence this is a subject which deserves more 

 thorough investigation. The evidence in some species is rather definite, 

 however. In the potato, for example, it is probable that bud mutations 

 are very, very rare. Yet the idea still prevails that disease resistant 

 strains of commercial varieties of potatoes can be obtained by hill selection 

 methods. In a variety susceptible to a given disease such strains would 

 have to originate as bud mutations and, while it is possible of course that 

 such a mutation in a given variety might occur, still in the light of what 

 is now known about the rarity of bud mutations in the potato it is prob- 

 able that in order to locate such a strain it would be necessary to test 

 millions of plants under conditions favorable for the disease. Severe 

 epidemics occasionally furnish opportunity for such selection on a grand 

 scale. But the scientific plan of procedure is to undertake variety test- 

 ing on a large scale preparatory to hybridization of the most promising 

 forms. 



The matter of adjustment between varieties and local environmental 

 conditions is of considerable importance. This is more widely recognized 

 in cotton perhaps than in any other crop. Cotton growers are generally 

 advised to secure locally grown seed, provided it has been properly selected 

 and handled. Unless recourse is had to the production of Fi hybrid 

 seed, this consideration of adaptation to local environment is destined to 

 become increasingly important as greater improvement is sought through 

 more intensive selection. For this reason seed production will probably 

 become more localized even though the business of handling and retailing 

 remains in the hands of comparatively few commercial establishments. 

 The recent rapid development and localization of truck crops in the 

 United States as reported by Blair is a case in point. Specialization 

 of this sort is bound to increase along with increase in population and 



