I04 PLANT-BREEDING 



ments of Nilsson at the Swedish Agricultural Experiment 

 Station at Svalof . Though working only in the interest of the 

 practical breeder, he applied to his selection thoroughly- 

 scientific methods and has arrived at the clear and unex- 

 pected conceptions of the variability of cereals and other 

 large agricultural crops, which we have exposed in our 

 previous chapters. Summing up their contents in a few 

 sentences, I have first to recall the practical results and 

 the numerous new and productive races, which have been 

 originated at Svalof and are now rapidly finding approval 

 with the agriculturists of Sweden, and even of Germany and 

 other countries. For scientific purposes they give proof of 

 the vaHdity of the methods employed at that station, and of 

 the accurate nature of the principles involved therein. 



Nilsson at first tried the usual German method, but 

 soon found that it yielded its results only in exceptional 

 cases and could not be applied to all the needs of the agri- 

 culturists (1885- 1 891). He then changed his principle and 

 sowed the kernels of numerous selected ears separately or 

 in small groups (1891-1892). The result was thoroughly 

 decisive, for all the parcels grown from mixed seeds gave 

 a mixed progeny, and only those which were derived from 

 one single ear each gave a pure and uniform culture. This 

 unexpected phenomenon was at once made the basis for 

 further experiments and in numerous sowings, where each 

 was derived from one single plant, the strains were almost 

 always found pure and constant. The only exceptions were 

 those in which a hybrid ear had been accidentally chosen. 

 Here of course the ordinary sphttings of hybrid progeny 

 were observed, but in choosing among their products, con- 

 stancy could be reached in many instances. 



Therefore Nilsson's principle for all breeding purposes 

 is now to derive his strains from single mother plants. Only 

 such strains give pure breeds. A second discovery made 



