194 ADAPTATION 



The Characins offer us the very best imagi- 

 nable proof, both for orthogenesis and against its 

 universality. The fact that Unes of evolution 

 radiate in so many directions in this family is ab- 

 solutely conclusive proof that there are many 

 possibihties, that evolution to adaptive points 

 may not only take place along one line or par- 

 allel hnes, but along very many diverging lines. 

 On the other hand, the fact that there are lines 

 with but few breaks leading from the general- 

 ized central type to such aberrant forms as the 

 minute sand-burrowing Characins, duplicating 

 our sand darters, or to the death-dealing Serra- 

 sahno, or the flying Gasteropelecus, shows that, a 

 path of adaptive modification once entered upon 

 by these fishes, evolution along that hne may 

 take place, even beyond the point of highest ad- 

 vantage. These hnes are not parallel and can 

 not therefore have been the result of the inherent 

 make-up of the family.^ They have in some way 

 been determined and are being followed to the 

 limit. 



E. Mutations. The possibihty of divergence 

 in many directions has been experimentally 

 demonstrated by De Vries, who, with others, has 

 claimed that the hne of adaptive modification is 

 broken, not bent. Waiving the question of 

 whether the difference between the bend and 

 break is one of kind or degree, permit me again 



' Similar characters like a pair of canines or ctenoid scales 

 have appeared in very diverse genera both in Africa and in 

 South America. 



