264 Edmund B. fVihon. 



ered inadequate for the cases of qualitative division reviewed 

 above. 



As regards the relation between self-differentiation and depen- 

 dent or correlative differentiation, our only guide must be the 

 indirect evidence derived from the response of the cell to the 

 change of conditions when its typical relation to the whole Is de- 

 stroyed by isolation or displacement from Its normal position. 

 For It Is perfectly obvious that If the "atypical" or secondary 

 changes characteristic of an isolated blastomere do not take place 

 In a complete embryo it is because of the relation of the cell to 

 the whole of which It forms a part; and It Is this "relation" that 

 renders the developing organism a unit, even in the most highly 

 differentiated type. As to what this "relation to the whole" 

 really Is we know practically nothing; but even though we employ 

 a phrase of vague and uncertain content it Is of use as Indicating 

 a unity or harmony of organization that is not destroyed by the 

 secondary distribution of the factors of differentiation among 

 localized centers. 



It is obvious that the differentiation even of such cells as the 

 primary trochoblasts, which possess so high a degree of self-dif- 

 ferentiation, must be definitely coordinated In some way with the 

 development of the embryo as a whole, as Is shown for Instance 

 by the remarkable manner in which the rows of cilia, at first dis- 

 connected, are ultimately fitted together to form continuous rings 

 In the prototroch; but It seems equally obvious that In such cases 

 corrflative differentiation subsequent to division plays but a minor 

 part In the Internal transformation of the cell. I may here point 

 out, however, that the lessened Inequality of division so frequently 

 observed In the Isolated blastomeres is posibly an indication of 

 regulative response on the part of the internal factors of cell-dif- 

 ferentiation. It is clear that the position of the spindle — and 

 hence the character of the ensuing division — is definitely cor- 

 related with the segregation-pattern ; and in the moUuscan egg 

 many, probably all, of the earlier unequal divisions are qualitative 

 In character. It is, therefore, a fair hypothesis that in these cases^ 



^The unequal division of teloblasts shows that the statement should not be 

 made general without further evidence. 



