234 SPECULATIVE SCIENCE 



unimportant ; ' or finally, lie may take refuge in the word 

 correlation, and say, other parts were useful, which by the law 

 of correlation could not exist without these parts ; and although 

 he may have not one single reason to allege in favour of any of 

 these statements, he may safely defy us to prove the negative, 

 that they are not true. The very same difficulty 'arises when a 

 disbeliever tries to point out the difficulty of believing that some 

 odd habit or complicated organ can have been useful before fully 

 developed. The believer who is at liberty to invent any ima- 

 ginary circumstances, will very generally be able to conceive 

 some series of transmutations answering his wants. 



He can invent trains of ancestors of whose existence there is 

 no evidence ; he can marshal hosts of equally imaginary foes ; 

 he can call up continents, floods, and peculiar atmospheres ; he 

 can dry up oceans, split islands, and parcel out eternity at will ; 

 surely with these advantages he must be a dull fellow if he can- 

 not scheme some series of animals and circumstances explaining 

 our assumed difficulty quite naturally. Feeling the difficulty of 

 dealing with adversaries who command so huge a domain of 

 fancy, we will abandon these arguments, and trust to those which 

 at least cannot be assailed by mere efforts of imagination. Our 

 arguments as to the efficiency of natural selection may be 

 summed up as follows : 



We must distinguish several kinds of conceivable variation 

 in individuals. 



First. We have the ordinary variations peculiar to each indi- 

 vidual. The effect of the struggle for life will be to keep the 

 stock in full vigour by selecting the animals which in the main 

 are strongest. When circumstances alter, one special organ may 

 become eminently advantageous, and then natural selection will 

 improve that organ. But this efficiency is limited to the cases 

 in which the same variation occurs in enormous numbers of 

 individuals, and in which the organ improved is already used 

 by the mass of the species. This case does not apply to the 

 appearance of new organs or habits. 



Secondly. We have abnormal variations called sports, which 

 may be supposed to introduce new organs or habits in rare in- 

 dividuals. This case must be again subdivided : we may sup- 

 pose the offspring of the sports to be intermediate between 



