X906] DE VRIES— ELEMENTARY SPECIES IN AGRICULTURE. 151 



as to respond to almost all the requirements of the agricultural 

 practice. By simply searching among them, the proper type may 

 forthwith be found for almost each gap in practice. In this way 

 they are seen to afford almost unexhaustible material for selection. 



For to-day's theme I have chosen an application of these discov- 

 eries of Nilsson to a criticism of the current views concerning the 

 bearing of agricultural breeding processes on the theory of evolution. 

 Formerly I gave the warning not to trust too much to these proc- 

 esses and to make use, in scientific discussions, only of the most 

 simple and clear cases (Mut. Theory, I, p. 59). The nev/ facts, 

 now at hand, go to prove that even the apparently simple methods 

 of selection have been far more complicated than their authors sus- 

 pected. The slow and gradual working up of a cereal to a previ- 

 ously fixed ideal seemed to be a process of the simplest possible 

 nature. In reality, however, it is composed of a series of factors, 

 which the breeders themselves have not recognized, and which, 

 therefore, it is now often impossible to discern in their descriptions. 

 In broad lines such an analysis has been made practicable by Nils- 

 son's discoveries. Unfortunately it conduces to a less high appre- 

 ciation of the breeder's merits (Mut. Th., p. 82), but on the other 

 hand it gives a stronger support to the theory of the saltatory origin 

 of species. 



The breeding of cereals results in varieties, which are as constant 

 and independent as the best horticultural sorts. In some cases they 

 are known to originate in the same way, by accidental sports, as in 

 the instance of Beseler's oats losing their needles. Here their com- 

 plying with the principle of mutation is obvious. In the large 

 majority of cases, however, including the most renowned improve- 

 ments of cereals and other crpps, they are said to have been pro- 

 duced by the common slow and gradual process of selection. All 

 such cases are surrounded with doubt, as well concerning their real 

 origin, as in view of the degree of self-dependency which is reached 

 at the end. Often practical reasons lead one to prefer the original 

 seed to one's own harvest, especially when it is difficult to keep the 

 cultures clean from vicinistic impurities. A race, which is really self- 

 dependent, may in this way seem to be permanently related to the 



PROC. AMER. PHIL. SOC, XLV. 183J, PRINTED OCT. 27, I906. 



