380 THE INFLUENCE OF LIVING SUKKOnNDINGS. 



firmed by actual observation of tbe development of one of the 

 most highly developed dorsal eyes of the Onchidiiim, coincide — 

 as I think must be allowed — ^with the phases of development 

 previously set forth on hypothetical grounds as being those 

 proper to the formation of an eye, if we. assume that by the 

 simple property of a regularly curved and smooth surface of a 

 papilla of the skin, all the light, heat, and chemical rays of a 

 beam of light must be made to converge in one point. 



If the foregoing chain of argument is correct, as I cannot 

 doubt, we have here detected an organ of extremely compHcated 

 structure in the very process of formation ; and this instance 

 proves, if indeed further proof were needed, that, as Darwin has 

 often insisted, an organ can never be created by natural selec- 

 tion, but can only be modified and improved by it. Sometimes, 

 no doubt, expressions are used, even by naturalists, which 

 might lead us to suppose that they were of opinion that an 

 organ might originate from its use — thus, in this case, an eye 

 from the act of sight ; this, of course, is absolutely eiToneous. 

 Sight, on the contrary, cannot occur till from other causes — 

 as in this instance the direct stimulus to the skin — those parts 

 have been formed which must be in existence before sight is in 

 any way possible. Here, in the dorsal eyes of Onchidium, in 

 the first instance it was the concentration of light on a certain 

 point within the skin, which was occasioned by the form and 

 structure of the papUlse, and the consequent modification of 

 certain cells, which gave rise to a primitive organ of sight, and 

 this organ was capable of still further development through 

 natural selection, since, from the very first, it contained within 

 itself the elements of further perfectibility and modification. 



It follows, moreover, that if at any time such a primitive eye 

 were developed in an animal which was not exposed to this 

 process of natural selection by its association with a fish that 

 preyed upon it, the organ, being useless, would disappear or at 

 any rate degenerate. And this actually is the case, as we have 

 seen ; for all the species of this genus which live where the 

 Periophthalmus does not, are devoid of these eyes. Only one as- 

 csrtaiaed exception is as yet known — the above-mentioned species 

 of Onchidium which lives associated with species that can see 



