( Ixxii ) 



The a priori considerations which appear to me to justify the 

 behef that correlation may exist between obviovis and concealed 

 characters, and that such correlation is therefore a worthy sub- 

 ject of investigation, are, in the present imperfect state of know- 

 ledge, necessarily of a very general character. The complexity 

 of the living organism regarded as a machine actively trans- 

 forming matter and energy, renders it almost certain that any 

 variation in a particular structure or function must be accom- 

 panied by a simultaneous modification of some other structure 

 or function. Moreover, since every external part of the 

 organism is more or less variable, it is a fair inference that 

 internal parts are also variable. On this point there is much 

 actual evidence with respect to the higher animals, but little 

 or nothing is known about the internal variability of the 

 " lower " animals, and this remark is particularly true with 

 regard to insects. Now, if internal structures are variable, 

 there is no reason why internal functions should not also be 

 variable. If, therefore, natural selection can take advantage 

 of variations of structure, I do not see how we can refuse to 

 admit that it can take advantage of variations of function 

 when such modifications are required in response to some 

 change in external conditions. It is to this point that I wish 

 particularly to direct attention on the present occasion, because 

 in view of the more obvious action of selection as an agency 

 capable of leading structural modifications in desired direc- 

 tions, it has not been sufhciently insisted upon that functional 

 variations must equally come under its influence. It is no 

 doubt due to the circumstance that variability of function is 

 so very much more difficult to deal with by observational or 

 experimental methods than variability of structure that the 

 contributions of morphology to the theory of descent have 

 hitherto far outweighed those of physiology. 



If it be admitted on such grounds as these that diagnostic 

 characters may not as such be of direct selection value, the 

 systematist will do well to pause before declaring the in- 

 adequacy of natural selection in any particular case. There 

 are many characters of direct and obvious utility, such as 

 those connected with the procuring of food, defence, con- 

 cealment, or other means of protection, secondary sexual 



