1 20 PLANT-BREEDING 



mentioned is of the same nature as that observed at Svalof 

 for the other cereals, and described in our previous lectures. 

 Thus the varieties are to be considered as built up of quite 

 numerous elementary forms, each of which is essentially 

 uniform and constant. The cross-pollination must, of 

 course, obscure this fact to some extent, but cannot annihi- 

 late it. As soon as such an elementary form is sufficiently 

 isolated and multiplied so that its progeny may fertilize it- 

 self exclusively, a uniform and constant race will be obtained. 

 Variability will then be limited to the smaller, but unavoid- 

 able changes, which climatic and environmental conditions 

 will always evoke, even in the most purely bred races. The 

 nature of these so-called fluctuations we shall soon have to 

 consider, but for selection they are only of secondary impor- 

 tance. 



The principle of selection at the Svalof Experiment 

 Station consists in the search for such elementary forms, 

 and in their isolation and subsequent comparative trial. 

 No purifying and no fixing, or in other words, no subsequent 

 or continuous selection is needed, provided the chosen ears 

 are not hybridized. Man cannot originate these variations, 

 nor can he essentially improve them. He must simply be 

 on the alert to recognize and isolate them and to compare 

 their progeny with the main strain. In the same way the 

 problem of corn-breeding is to recognize these elementary 

 races. All success depends upon finding the best among 

 them and on thus taking complete advantage of the variabil- 

 ity already existing in the fields. Even the races with spe- 

 cial characters, as, for instance, those with a high yield of 

 oil or of protein, have, as a fact, been secured in this same 

 way. 



Corn, however, differs from the other cereals in some 

 very important points. Two of them are now to be consid- 

 ered. One is the open pollinized condition, and the other 



