410 W. Bateson. 



the ground lias been far more Completel}^ surveyed, I am lotli to 

 attempt an answer. De Vries has defined somewhat strictly the 

 diiferences betwen specific and varietal distinctions, asserting that it is 

 the latter alone which exhibit Mendelian heredity (115, 116). Though 

 I agree with him in perceiving- that genetic research may ultimately 

 provide some approach to a valid distinction between species and 

 variety, I am reluctant to accept any evidence yet attained as an 

 adequate basis for so vast a generalisation. Of the wow-Mendelian 

 phenomena of heredity we know as yet almost nothing. I believe 

 that the consequences of specific crosses — in the stricter sense — 

 have thus far in no single case been fully explored. Even of the 

 results of crosses between '■^petites espèces'' differing in several char- 

 acters, little is known. 



The phenomenon of sterility certainly counts for much in this 

 part of genetics. As to the nature of this sterility and its limitations, 

 even as to the rules of inheritance in those cases where sterility is 

 partial, we have scarcely any adequate knowledge, yet such know- 

 ledge must be obtained as a necessary preliminary to a reliable 

 judgment on the limitations of specific distinction. All this field of 

 research is open now, and before long we shall know what it contains. 

 From the little however that we do already know, I find it im- 

 possible to draw the conclusion to which de Vries would commit us. 

 Surely, for instance, the long range of partially sterile derivatives 

 from the various crosses of Narcissus species show abundant signs of 

 segregation. To deny that N. pseudo-narcissus and poeticus for in- 

 stance, are species in the strictest sense is impossible. They are the 

 parents of many of these semi-sterile forms. Similar cases will occur 

 to the reader in abundance. De Vries himself perhaps is feeling this 

 difficulty when [3Intationstheone II, p. 646) he speaks oî Lychnis vespertina 

 and diiirna as being distinguished partly by varietal and partly by 

 specific differences. We may indeed, as I have said elsewhere 

 (R.E.C. (10), p. 148) be driven to conceive specific difference as a 

 property of the residue or basis upon which the allelomorphic char- 

 acters are implanted; but it is not easy to suppose that the features, 

 breadth of leaves, and length of flowering stem — named by de Vries 

 as non -splitting characters in Lychnis — are of this fundamental 

 nature. 



Feeling thus the impossibility of now defining the segregating 

 from the non- segregating, I am unable to follow de Vries in the 

 further step which he has taken (115). in assigning a definite physio- 

 logical reason for the difference between these classes. His suggestion 

 is that in Mendelian heredity there is a process spoken of as "Bi- 

 sexual", in which each determining factor (Anlage) meets a corre- 

 sponding opponent; while in the other or "unisexual" case, the factor 



