HYBRIDIZATION AND MUTATION 413 



accordance with Mendelian principles. It cannot be doubted that 

 hybridization occurs occasionally in cultivated cereals, and 

 Professor de Vries is of opinion that the occurrence of the 

 different types or mutations is often the result of hybridization 

 at various periods in the history of the race : 



"Experience, however, shows that in ordinary fields almost 

 all possible combinations may be met with, and it is to be pre- 

 sumed that at least the greater number of them are due to 

 crosses in previous and, perhaps, in long-forgotten years." x 



Some of these combinations, as might be expected, are not 

 stable but split up into a number of varieties in the next 

 generation, but also : 



" We may conclude that some, and perhaps many, of the types 

 which may be selected and isolated in the fields and which prove 

 to be constant races must be of hybrid origin." 2 



De Vries maintains that in the case of the cereals so many of 

 these " types " now lie ready to our hand that all we have to do 

 is to pick out those which we require and cultivate them in 

 isolation from each other and from the remainder. 3 Professor 

 Biffen has shown, however, as we have already pointed out in 

 Chapter XIV, that it is possible by intelligent artificial hybridi- 

 zation to produce yet other stable combinations or hybrids 

 which may surpass in value any which have accidentally arisen 

 in the past. 



It appears, then, that many at any rate of the so-called 

 mutations or types amongst cereals are due to hybridization. 

 How far this applies to mutations in general it is quite impossible 

 to decide. That it is not always so, however, appears to be proved 

 by the occurrence of such mutations or sports as hexadactylism, 

 which are known to be inherited and which cannot have 

 arisen in this way. 



De Vries tells us in another work that : 



" According to the theory of mutation species have not arisen 

 gradually as the result of selection operating for hundreds, or 

 thousands, of years but discontinuously by sudden, however 

 small, changes. In contradistinction to fluctuating variations 



1 De Vries, i>2>. cit., p. 80. 



2 Ibid., loc. cit. 



3 1 bid., p. 50. It seems strange, considering that de Vries admits that many 

 at any rate of the "types' 1 have probably arisen by hybridization in the first 

 instance, that he should attribute so little value to artificial hybridization as a 

 means of improvement. 



