224 SPECULATIVE SCIENCE 



with very ordinary beasts and species. They should select the 

 descendants which might be rather better in the first generation, 

 and then should carefully abstain from all attempts at improve- 

 ment for twenty, thirty, or one hundred generations. Then 

 they might take a little step forward, and in this way in time 

 they or their children's children would obtain breeds far surpass- 

 ing those produced by their over-hasty competitors, who would 

 be brought to a stand by limits which would never be felt or 

 perceived by the followers of the maxim, Festina lente. If we 

 are told that the time during which a breeder or his descen- 

 dants could afford to wait bears no proportion to the time used 

 by natural selection, we may answer that we do not expect the 

 enormous variability supposed to be given by natural selection, 

 but that we do expect to observe some step in that direction, to 

 find that by carefully approaching our limit by slow degrees, 

 that limit would be removed a little farther off. Does anyone 

 think this would be the case? 



There is indeed one view upon which it would seem natural 

 to believe that the tendency to revert may diminish. If the 

 peculiarities of an animal's structure are simply determined by 

 inheritance, and not by any law of growth, and if the child is 

 more likely to resemble its father than its grandfather,, its 

 grandfather than its great-grandfather, &c., then the chances 

 that an animal will revert to the likeness of an ancestor a 

 thousand generations back will be slender. This is perhaps 

 Darwin's view. It depends on the assumption that there is no 

 typical or average animal, no sphere of variation, with centre 

 and limits, and cannot be made use of to prove that assumption. 

 The opposing view is that of a race maintained by a continual 

 force in an abnormal condition, and returning to that condition 

 so soon as the force is removed ; returning not suddenly, but by 

 similar steps with those by which it first left the average state, 

 restrained by the tendency to resemble its immediate progenitors. 

 A priori, perhaps, one view is as probable as the other ; or in 

 other words, as we are ignorant of the reasons why atoms fashion 

 themselves into bears and squirrels, one fancy is as likely to 

 meet with approval as another. Experiments conducted in a 

 limited time point, as already said, to a limit, with a tendency 

 to revert. And while admitting that the tendency to revert 



