370 GEORGE JOHN ROMANES 1894 



at the origin, and throughout the growth of every 

 branch? Moreover, I agree with you about self- 

 fertilisation, but see in it a form of physiological 

 selection ; it is one kind of sexual isolation, or 

 prevention of inter-crossing with neighbouring in- 

 dividuals. So that the more perfectly it obtains in 

 any given type, the better chance there is for that 

 type to become a new species by independent varia- 

 bility and this whether or not the independent 

 variability is likewise indiscriminate (or in your 

 terminology ' indefinite '). 



In my last letter I referred to the works of Jordan 

 and Nageli for any number of 'facts in Nature of 

 varieties arising among the type-forms.' I will show 

 you the passages when we meet. But even in cases 

 of ' local varieties,' where a variety has a habitat 

 of its own surrounded by the type-form, I should 

 expect experiment would often (though by no means 

 always) show some degree of cross-infertility between 

 the two, pointing to prepotency (i.e. early stages of 

 physiological selection) being the origin of the diver- 

 gence. 



Before we meet I wish you would try to think of 

 any plants which can be propagated by cuttings (or 

 otherwise asexually) which are known to be modifi- 

 able by changed conditions of life in the first genera- 

 tion. I understand you that in some cases the seed 

 of such a plant will not revert when sown in its 

 natural environment, though, of course, the rule is 

 that it does. Well, in either case, I should much 

 like to try whether a cutting &c. from the trans- 

 planted (and therefore modified) tubers &c. would 



