HEREDITY, VARIATION 107 



manifested in Socrates and Plato, Homer and Virgil, 

 Alexander and Phidias, Buddha and Confucius in the older 

 world ; in Shakespeare and Newton, Michael Angelo, Fara- 

 day, and Darwin in more recent times. 



And with all this endless variation wherever we look for 

 it, we are told again and again in frequent reiteration, that 

 variation is minute, is even infinitesimal, and only occurs at 

 long intervals in single individuals, and that it is quite in- 

 sufficient for natural selection to work with in the production 

 of new species. 



This blindness, no doubt, arose in some persons from 

 the ingrained idea of man's special creation, at all events, 

 and that it was almost impious to suppose that these 

 variations could have had anything to do with his develop- 

 ment from some lower forms. But among naturalists the 

 idea long prevailed, as it does still to some extent, that in a 

 state of nature there is little variation. Yet here, too, they 

 might have found a clue in the fact, so often quoted, that 

 a shepherd knows every individual sheep in his flock, and 

 the huntsman every dog in his well -matched pack of 

 hounds, and this notwithstanding that in both cases these 

 animals are selected breeds in which all large deviations 

 from the type form are usually rejected. 



Of late years, however, variations occurring in a state 

 of nature have been carefully examined and measured, and 

 it is to some of these that we will now appeal for the proof 

 of ever-present variation of the character and amount 

 needed for the production of new species and of every kind 

 of adaptation by means of natural selection or the survival 

 of the fittest. Before giving examples of the variation of 

 the higher animals it will be advisable to show what is 

 meant by the " law of frequency " of variations which has 

 been established by the measurement of several thousands 

 of men in various countries of Europe. These when 

 recorded by means of a diagram are found to form a 

 very regular curve, which becomes more and more regular 

 the larger are the numbers measured. The importance of 

 this is that when we have only small numbers of animals to 

 deal with, and we find great irregularity in their diagrams, 

 we are sure that if we had measurements of hundreds or 



