ELEMENTARY SPECIES IN AGRICULTURE. 427 



ments, at the same time indicated the way for an explication on 

 the basis of our present views concerning the different types of 

 variabiHty. 



It would take me too long to describe the methods and cultures 

 of the Minnesota Experiment Station, and I may assume that their 

 leading principles and practical results are well known. But I wish 

 to point out that, exactly in the principle of sowing the seeds of 

 individual selected plants separately, Hays gained a distinct advan- 

 tage over the slow process of Rimpau and the other German breeders. 

 He found, by his method, that the isolated strains are at once con- 

 stant and pure. They had only to be multiplied in order to give a 

 new race. Of course, the different mother plants had to be com- 

 pared in their progeny, and among a large number of such new 

 pedigree races only one or two were found to be the very best. The 

 remainder had to be rejected, and only those few excelling ones could 

 be introduced with advantage into the field-cultures of the state. 



If now we compare this principle of Nilsson and Hays with the 

 method of Rimpau we find that the Swedish and American breeders 

 by one single choice isolated the very best strains and observed them 

 to be constant and pure. The German breeders, on the other hand, 

 by selecting a number of ears, must have got impure races, and 

 wanted a long succession of years and a constantly repeated selection 

 in order to reach the same result in the end. 



Hence we may deduce the supposition that if Rimpau in starting 

 his experiments, now forty years ago, had had at his disposition our 

 present knowledge of variability, he would have sown the kernels 

 of his selected ears separately and selected at once among the resul- 

 ting strains the very one which now bears the name of his farm. No 

 continuous culture and repeated selection would have been needed, 

 and the false appearance of a slow and gradual improvement of a 

 race by selection would simply have been avoided. 



The German breeding process has always been one of the most 

 valuable arguments for the theory of gradual selection and was of 

 late considered as its last botanical support. By means of the dis- 

 coveries of Nilsson and Hays this support has now been broken 

 down, and agricultural selection is no longer an argument against 

 the conception of an origin of species by saltatory changes. 



(Proceedings of the American Philosophical Society, 

 Vol. XLV, igo6, p. 149.) 



