72 THE BUSTS OF GBAIXS IN THE UNITED STATES. 



regard to yellow rust, he found that on crossing susceptible and 

 resistant varieties the offspring was susceptible. Upon self-fertiliza- 

 tion of these susceptible individuals, resistant and susceptible descend- 

 ants were produced in the proportion of one of the former to three of 

 the latter, that is, resistance was recessive to susceptibility, the 

 degree of susceptibility being variable. When the degree of suscep- 

 tibility differed in the two parents the hybrid resembled the more 

 susceptible parent in that respect. More important still, the rela- 

 tively resistant forms bred true to these characters in the succeeding 

 generations. Bolley (27, pp. 182, 183), from several years' work in 

 the selection and breeding of flax and wheat resistant to wilt and 

 rust, respectively, came to similar conclusions, and in addition 

 believes that unit characters of resistance may be originated even 

 from a very susceptible variety by gradually subjecting the crop to 

 disease from year to year. He maintains that these characters may 

 later be inheritable. 



The authors have been engaged in similar work since 1907, but 

 sufficient results have not yet been obtained to pronounce definitely 

 on the question of the application of Mendelian laws to resistance to 

 rust in these experiments. Detailed results of this work are reserved 

 for future publication. 



METHODS USED IN SELECTION AND BREEDING. 



From the foregoing it will be seen that there are three methods in 

 use for the development of rust-resistant grains through selection and 

 breeding. (1) A careful testing and selection of pure varieties to 

 determine which are already resistant; (2) selection of the best 

 individuals or bulk selection from some strain or variety from year 

 to year under fairly constant disease conditions in the belief that 

 disease resistance is accumulative; (3) hybridizing of desirable 

 varieties with some variety of known resistance and selecting the 

 resistant plants. 



The first method is absolutely necessary before the third can be 

 applied, while the second is possible for any worker along this line at 

 any time. 



In breeding for resistance to almost any disease, in order to insure 

 rapid progress, the disease must be present every year in sufficient 

 virulence to affect the crop under trial with more or less severity. Cer- 

 tain diseases, particularly rust, occur in epidemical proportions only 

 at irregular intervals. This not only delays results in nonepidemical 

 years but disturbs them in other ways. To overcome these objections, 

 diseases must be promoted yearly on the breeding grounds in every 

 possible way. In order to do this, special breeding plats are employed 

 If one is working for resistance to flax wilt, the breeding plat must be 

 on flax-sick soil; if for drought resistance, on ground particularly 



216 



