Conclusion. 509 



Inconstant species^ do not seem to occur in nature. 

 And if they did, they would be bound to disappear sooner 

 or later because the atavists, which they produce every 

 year, would probably be constant, and soon supplant the 

 inconstant form. 



We see therefore that in the process of the origin of 

 new species some certainly do arise which are not capable 

 of existence for any length of time. Nature does not 

 confine herself to producing just what is wanted; her 

 creative power seems to be almost unlimited. She fur- 

 nishes every possibility, so to speak, and leaves it to the 

 environment to choose what suits it. In other words 

 mutability is indiscriminate. 



Could the new species maintain themselves in nature ? 

 This is a question which naturally presents itself. I have 

 conducted no experiments in this direction. But it is a 

 fact that 0. laevifolia and 0. brevistylis have survived the 

 struggle for existence at Hilversum for a long period of 

 time.^ In my garden the plants certainly appear to be 

 less resistent than they would be in nature, partly on 

 account of the amovtnt of manure they get, partly because 

 there is no selection which kills off the delicate ones in 

 early youth, and partly and in fact chiefly, on account of 

 my preference for annuals. For example annual ohlonga 

 set far too little seed ; biennial ones give a good harvest. 

 O. rubrinervis as an annual is very brittle, as a biennial 

 very strong; and so forth. Both these and O. gigas 

 would maintain themselves Avell in nature if they grew 

 in the biennial form; and would perhaps form just as 

 good species as the forms imported from America, O. 



*For the justification for this expression see § 19, PP- 377-379- 



^They are still seen to be growing in the field at Hilversum 

 (Note of 1908.) 



