778 LOUIS AGASSIZ. 



one was completed. It was his last word 

 upon science ; the correction of the proof- 

 sheets was the last act of his working life, and 

 the article was published after his death. In 

 it he claimed that the law of evolution, in a 

 certain sense as true to him as to any so-called 

 evolutionist, was a law " controlling develop- 

 ment, and keeping types within appointed cy- 

 cles of growth." He maintained that this law 

 acts within definite limits, and never infringes 

 upon the great types, each one of which is, in 

 his view, a structural unit in itself. Even met- 

 amorphoses, he adds, " have all the constancy 

 and invariability of other modes of embryonic 

 growth, and have never been known to lead 

 to any transition of one species into another." 

 Of heredity he says : " The whole subject of 

 inheritance is exceedingly intricate, working 

 often in a seemingly capricious and fitful way. 

 Qualities, both good and bad, are dropped as 

 well as acquired, and the process ends some- 

 times in the degradation of the type, and the 

 survival of the unfit rather than the fittest. 

 The most trifling and fantastic tricks of inher- 

 itance are quoted in support of the transmuta- 

 tion theory; but little is said of the sudden 

 apparition of powerful original qualities, which 

 almost always rise like pure creations, and are 



