60 THE WANDERINGS OF ANIMALS [OH. 



to be picked out by natural selection, and further 

 developed by cumulation, it would follow that each 

 ancestral group possessed a whole arsenal of every 

 kind of variation, ready to be selected when occasion 

 required it. Descendants of the same frog, which 

 found themselves in a forest, drew upon its incipient 

 finger-disk variations, be these ever so tiny ; others, 

 in inundated country, improved upon the webbed 

 propensities, and yet others encouraged those varia- 

 tions which assured a chance of digging feet. This 

 is of course, to a certain extent, our idea of a ' general- 

 ised ancestor,' some kind of anurous Amphibian, for 

 all the toads, frogs and tree-frogs. 



This view implies that by this time the arsenal 

 of variations has been exhausted, there being in 

 forests only forest creatures, in deserts only desert 

 specialists, all of which have of course lost the other 

 non-fitting characters by the process of weeding out. 

 Consequently any further change in the environment 

 would kill the present faunas. What was easy for an 

 indifferent organism, must have become increasingly 

 difficult for the specialist, and an all-round speciali- 

 sation in every direction is impossible. This is covered 

 to a great extent by Cope's law of the over-specialised, 

 and by D. Rosa's ' progressive reduction of variability/ 

 We cannot here enter into a discussion of these 

 important topics. Suffice it to say that variability is 

 not the same as the incessant cropping up of slight, 



