THE ROYAL WALNUT 79 



reason to doubt that the dwarf form and the rela- 

 tively large one are descended from the same 

 original stock, though doubtless divergence has 

 gone on through numberless generations. 



Meantime the English or Persian walnut, the 

 other parent of the Paradox, is also a variable 

 tree. In its native home it is very small, and 

 even the cultivated variety cannot be depended 

 upon to reproduce a given racial strain when 

 grown from the seed. 



It is obvious, then, that the tendency to dwarf- 

 ness, which appears in such conspicuous fashion 

 in some of our second generation hybrids, may 

 be accounted for as reversion to dwarfed ances- 

 tral strains in both parents in the case of the 

 Paradox and of one parent in the case of the 

 Royal. 



The tendency to grow relatively large pre- 

 vailed in the strains of walnuts that were used in 

 my hybridizing experiments, and the prepotency 

 or dominance of this tendency is clearly shown in 

 the hybrids of the first filial generation. But the 

 latent tendency to dwarfness, which in the Men- 

 delian phraseology would be termed a recessive 

 trait, is able to reassert itself in a certain number 

 of the offspring of the second filial generation, 

 causing these to "throw back" to their dwarfed 

 ancestors in the fullest measure. 



