374 
De Vries on the Orighi of Sjyecies 
completely excludes any such remarkable form of regression as that which Professor 
de Vries describes but fails to demonstrate among the offspring of his " mutations." 
I feel confident that when this result is better understood than it is at present 
such naturalists as Professor de Vries and Mr Bateson will abandon their attempts 
to distinguish between " variations " and " mutations," or between " normal " and 
" differontiant " variations. These attempts appear always to rest upon a fancied 
relation between the phenomenon of " regression '' and the stability of specific 
mean character through a series of generations which a little knowledge of the 
statistical theory of regression will show to be wholly imaginary. 
