54 COSMIC PHILOSOPHY. [pt. il 



are linked together only at their lowest extremities, and 

 wliile the lowest vertebrate has an eye far inferior to the 

 one just described, the molluscoid ascidians have merely 

 rudimentary eye-spots. The coincident structures have there- 

 fore been independently developed. Again, Mr. Mivart urges 

 that the agreement cannot be explained on the assumption 

 " that the conditions requisite for effecting vision are so rigid 

 that similar results in all cases must be independently arrived 

 at " ; for the eyes of the higher insects, which are excellent 

 visual organs, differ very widely in structure from those of 

 the cuttle-fish and tlie higher vertebrates. Here, therefore, is 

 a difficulty ; and it is still further increased if the alleged fact 

 be true, that there is a similarly close correspondence between 

 the auditory structures in the vertebrates and in the cuttle-fish. 



In presenting these difficulties I have closely followed Mr. 

 Mivart, whose scientific arguments are usually stated with a 

 clearness and precision which one would gladly see paralleled 

 in the philosophic discussions by which they are supplemented. 

 I have selected these arguments because they seem to me to 

 constitute the strongest portion of the case which j\Ir. Mivart 

 has brought to bear against the theory of natural selection ; 

 and also because by seeing whither they tend, we shall begin 

 to see how the theory of natural selection must be supple- 

 mented, before it can become a complete explanation of the 

 phenomena with which it deals. 



Now we must at the outset admit that natural selection 

 must act upon every individual variation which is distinctly 

 advantageous or injurious to the species, — always preserving 

 the former and rejecting the latter. This process must equally 

 go on, whether the variation is a mere idiosyncrasy, such aa 

 we call fortuitous, or whether it is one that is manifested 

 simultaneously by a large number of individuals, so that it 

 may be traced to causes acting upon them all in common. 

 Now this latter ca^e is the one which must here be taken into 

 the account. If a large number of individuals may simul- 



