1894. NEUTER INSECTS AND LAMARCKISM. 95 



may it not in some cases raise the general level of efficiency, and so 

 improve neuter insects independently of use-inheritance ? 



Those who hold that Natural Selection can preserve highly 

 complicated efficiency by fully counteracting the lowering power of 

 disuse (and underfeeding?) in innumerable co-ordinated points — and 

 yet that it can do no more in the case of higher animals where no 

 such opposition has to be overcome — must face the logical conclusion 

 from their own premises, that the actual power of use-inheritance is 

 precisely nothing — Natural Selection being exactly as powerful either 

 with or without its opposition. 



The cherished opinion that complex evolution is impossible 

 without the aid of use-inheritance, is by no means widely accepted. 

 Darwin and Wallace, co-discoverers of the principle of Natural 

 Selection, have deliberately rejected the assumption. It seems, 

 indeed, to be merely one of those artificial difficulties which 

 philosophers in all ages have been wont to construct for themselves — 

 much as Leibnitz convinced himself that the law of gravitation was 

 impossible. The process of complex evolution may be compared to 

 a series of competitive examinations, in which failure in any one of 

 many subjects is fatal to success. Surely the standard of general 

 efficiency must be raised by the rejection of the least efficient candi- 

 dates. Otherwise, what is the use of competitive examinations ? 

 Selection, whether artificial or natural, cannot well be neutral in its 

 effects. Evidently successful or selected candidates will display 

 more than average capacity for carrying out the combined work 

 exacted of them by the examiners. Surely a similar process in nature 

 must similarly raise the level of the miniina in all essential points or 

 subjects without correspondingly preventing a raising of the maxima, 

 which, of course, will often vary upwards, since organic variability 

 is in all directions. Let us take another comparison. The strength 

 of a chain is that of its weakest link. If chains with the weakest 

 links are rejected, the chains that survive the test will manifestly be 

 stronger than those that break. Rigorous elimination of weak links, 

 without any correspondingly rigorous rejection of strong links, must 

 improve the strength or fitness of the links in general, however 

 numerous they may be. Organisms are such chains ; and so long as 

 defective links or viinima are eliminated more often than superfluously 

 strong links or maxima, the average of efficiency in all the links may 

 well be a rising one. If defective parts are incessantly eliminated 

 (together, of course, with the organisms which they ruin) Natural 

 Selection may gradually raise the standard of efficiency in all the 

 necessary co-operative parts or organs. Such a process cannot fairly 

 be described as an oft-repeated " fortuitous concourse of variations " ; 

 for all that is needed is ordinary variations and their extinction, so 

 far as they are hostile to general efficiency and success. 



Mr. Spencer at one time represented the chances against complex 

 evolution by Natural Selection as " infinity to one." He has since 



