316 SCIENCE PROGRESS. 



the considerations stated above we should expect that the 

 changes in chemical composition of the cytoplasm and the 

 correlated changes in the nucleus, in other words the dif- 

 ferentiation, would first become manifest in the peripheral 

 parts of the growing ccenocyte, and that we should have a 

 stage in which there was a cellular external layer and a 

 ccenocytial internal mass. We find that in fact in the 

 embryoes of many Ccelenterates the outer layer is divided 

 up early into sharply defined cells at an early period, whilst 

 the central cells retain the character of a ccenocytium ; at 

 most the cell outlines of the internal mass are confused and 

 indistinct. 



We see also that in the growing tissues of the embryoes 

 of higher animals the embryonic tissue is not cellular but is 

 a ccenocytium, for example, the mesoblast of Avian and 

 Selachian embryoes and of the Rabbit. It is only at a later 

 stage when different relations to other parts of the body 

 have been acquired and new exchanges of material are 

 forced upon the growing mass, that the continuous mass of 

 cytoplasm is split up into corpuscles, each of which, in my 

 view, corresponds to the limit of influence of a nucleus. 



On the other hand we have the undoubted fact that in 

 many organisms there is no ccenocytial phase in develop- 

 ment, but the cytoplasm surrounding the nuclei as they are 

 successively formed is immediately marked off into definite 

 corpuscles, so that the whole process of development 

 suggests the formation of an aggregate of bionts derived by 

 division from a single parental biont. An explanation of 

 this fact presents many difficulties, and I have not now 

 the space to discuss these difficulties and to show that, 

 obscure as the subject still is, there is ground for supposing 

 that the difficulties are chiefly due to the prepossession 

 which exists in most minds in favour of the independent 

 life unit theory. I hinted in my previous paper (loc. ciL, p. 

 171) that the discrete condition of the blastomeres of so 

 many embryoes may be connected with the fact that they 

 are, from the very outset, specialised. This means that as 

 the nucleus is in some way associated with the transmission 

 of historic qualities, these qualities may be located in special 



