INHERITANCE OF ACQUIRED CIIARACTERb 203 



but in (Jcfiiult of other ideiu'c for tlio inhoritjincc of experi- 

 ences, the exj)hinati()n is not satisfactory. But it is not easy 

 to bcHcve that in early times those witliout this instinct fell 

 victims to venomous snakes througii tlieir own fearlessness. It 

 is perhaps not necessary to take sides on this question. Any 

 view we may adopt rests for confirmation mainly on the im- 

 probability of what we conceive to be the opposite alternative. 

 Conklin further observes that the 



"so-called facts [of progressive heredity] are merely probabilities of a 

 higher or lower order, and to one man they seem more important than 

 to another. No conviction based even upon a high degree of proba- 

 bility can ever be reached in tliis way. There is here a deadlock of 

 opinion, each challenging the other to produce indubitable ]>r<)()f. 

 This can never be furnished by observation alone. Possibly even 

 experiment may fail in it, but at least it is the only hope." 



We shall not assist science, says Osl^orn, 



"wdth any evolution factor grounded uj)on logic rather than upon 

 inductive demonstration. A retrograde chapter in the history of sci- 

 ence would open if we do so and should accept as established laws 

 those which rest so largely upon negative reasoning." 



Meanwhile we may regard the theory of the inheritance of 

 acquired characters as a piece of useful scaffolding which has 

 served its punpose in the development of the facts of the deriva- 

 tion of species. At present most of it — perhaps all of it — must 

 be taken down, but it may be that from the same base will arise 

 a better constructed theory which will again serve a ])urpose 

 in the study of organic evolution. 



Similar conditions in life tend to develop or encourage 

 analogous adaptations in groups of animals not homologous in 

 structure nor closely related by lines of descent. In many cases 

 these adaptations are so very similar and are so subtle in their 

 parallelism as to deceive even the trained naturalist. In other 

 cases, the convergence and its consetpient analogies are less 

 perfect, and the separate influence of like selection und(M- like 

 environment is easily traceable. 



Examples of this sort are seen in the density of the fur of all 

 Arctic animals, whatever the group to which they belong. An- 

 other illustration is found in tbc whit^^ winter dress of wcuseb, 



