272 PLANT-BREEDING 



to point out the way in which further progress might be 

 attained. 



At the Swedish agricultural experiment station of Svalof, 

 however, it soon became evident, that this principle might 

 some day become the basis of exact methodical work. The 

 fact once ascertained, that the fields of our ordinary crops 

 are by no means uniform, and moreover conceal a noticeable 

 number of excellent types, suitable for the most widely 

 different requirements, the need was instantly felt for a 

 method which would enable the breeder to make his selec- 

 tions as large and as profitable as possible. According to 

 the rules already explained, this choice must always be a 

 two-fold one. In the first place comes the work in the 

 ordinary fields, the picking out of the aberrant individuals 

 which seem to promise some special result. Thousands of 

 ears are inspected and tested, and on account of certain 

 marks those are chosen which do not comply with the main 

 type. Then their seeds are sown, and the progeny of each 

 single specimen is studied and compared with the already 

 cultivated varieties. This comparison depends in part on 

 the same marks as the initial choice and in part on the 

 direct control of the agricultural quahties. By means of 

 this latter the judgment is finally made independent of the 

 laws of correlation, which by this contrivance afford only 

 the means to reach the aim by a far shorter and easier 

 way than would be possible if the first choice itself had to 

 be made on the basis of exact measurements of the practical 

 value of the individual plants. 



We thus see that we may hmit our study to the initial 

 choice. We may leave the subsequent work out of consider- 

 ation, since, as far as the correlations are concerned, they 

 require the appreciation of quite the same marks. But in 

 order to make the first choice as reHable and as profitable 

 as possible, a thorough knowledge is required. It must 



