7 8 ON HEREDITY. [II. 



phenomena of heredity by means of a hypothesis which corre- 

 sponds to a considerable extent with that just described. If we 

 substitute gemmules for molecules we have the fundamental 

 idea of Darwin's provisional hypothesis of pangenesis. Particles 

 of an excessively minute size are continually given off from all 

 the cells of the body; these particles collect in the reproductive 

 cells, and hence any change arising in the organism, at any 

 time during its life, is represented in the reproductive cell ^ 

 Darwin believed that he had by this means rendered the trans- 

 mission of acquired characters intelligible, a conception which 

 he held to be necessary in order to explain the development 

 of species. He himself pointed out that the hypothesis was 

 merely provisional, and that it was only an expression of 

 immediate, and by no means satisfactory knowledge of these 

 phenomena. 



It is always dangerous to invoke some entirely new force in 

 order to understand phenomena which cannot be readily ex- 

 plained by the forces which are already known. 



I believe that an explanation can in this case be reached by 

 an appeal to known forces, if we suppose that characters ac- 

 quired (in the true sense of the term) by the parent cannot 

 appear in the course of the development of the offspring, but 

 that all the characters exhibited by the latter are due to primary 

 changes in the germ. 



This supposition can obviously be made with regard to the 

 above-mentioned colony with its constituent elements differen- 

 tiated into somatic and reproductive cells. It is conceivable 

 that the differentiation of the somatic cells was not primarily 

 caused by a change in their own structure, but that it was 

 prepared for by changes in the molecular structure of the 

 reproductive cell from which the colony arose. 



The generally received idea assumes that changes in the 

 external conditions can, in connection with natural selection, 

 call forth persistent changes in an organism ; and if this view 

 be accepted, it must be as true of all Metazoa as it is of uni- 

 cellular or of homogeneous multicellular organisms. .Supposing 

 that the hypothetical colonies, which were at first entirely made 

 up of similar cells, were to gain some advantages, if in the 



* See Danvin, ' The Variation of Animals and Plants under Domes- 

 tication,' 1875, vol. ii. chapter xxvii. pp. 349-399. 



