ROMANES' MENTAL EVOLUTION, ETC. Sf 



selection must be always more or less objectionable, 

 as too highly charged with metaphor for purposes of 

 science, there is nevertheless a natural selection which 

 is open to no other objection than this, and which, 

 when its metaphorical character is borne well in mind, 

 may be used without serious risk of error, whereas natu- 

 ral selection from variations that are mainly fortuitous 

 is chimerical as well as metaphorical. Both writers 

 speak of natural selection as though there could not pos- 

 sibly be any selection in the coutse of nature, or natural 

 survival, of any but accidental variations. Thus Mr. 

 Eomanes says : * " The swamping effect of free inter- 

 crossing upon an individual variation constitutes per- 

 haps the most formidable difficulty with which the 

 theory of natural selection is beset." And the writer 

 of the article in the Times above referred to says : 

 " In truth the theory of natural selection presents many 

 facts and results which ' increase rather than diminish 

 the difficulty of accounting for the existence of species." 

 The assertion made in each case is true if the Chatles- 

 Darwinian selection from fortuitous variations is in- 

 tended, but it does not hold good if the selection is 

 supposed to be made from variations under which 

 there lies a general principle of wide and abiding 

 application. It is not likely that a man of M^. 

 Eomanes' antecedents shotild not be perfectly awake 

 to considerations so obvious as the foregoing, and I 

 am afraid I am inclined to consider his whole sugges- 

 tion as only an attempt upon the part of the wearer 



* Nature, August 5, 1S86. 



