314 C. M. CHILD 



from posterior portions of the body. In the posterior region of 

 the first zooid, however, the correlation with posterior regions is 

 necessary, under the usual conditions, for the formation of a head. 

 But when a head has once formed at any level of the original 

 body "we find that the capacity of the adjoining regions for head 

 formation has been correlatively increased. In general we may 

 say that in regions adjoining a previously existing head the process 

 of head formation after isolation approaches more or less closely 

 to 'self-differentiation,' or as I prefer to call it, 'constitutional 

 differentiation,' while the correlative factor in head formation 

 increases in importance with increasing distance of the level from 

 the previously existing head. Similar relations obtain in the 

 formation of posterior ends: in the posterior region of the first 

 zooid the formation of the tail is much more nearly a constitu- 

 tional process, while in pieces from the anterior end of this zooid 

 it is to a much greater extent a process of correlative differenti- 

 ation. These differences mean, I believe, that the regions adjoining 

 a head possess a certain constitution in consequence of their posi- 

 tion in the body, i. e., in consequence of the correlative effects to 

 which they are subjected. This constitution is such that all that is 

 necessary for the formation of a head by these parts is the isolation 

 of a sufficiently large mass from the previously existing head, 

 irrespective of whether posterior regions are present or not. 

 If these conclusions are correct then it becomes less difficult to 

 understand why the formation of a new head at a level far pos- 

 terior to the original head can occur only when a considerable 

 region posterior to it is present. Before section this level was 

 in no way closely associated with the old head and the effect of 

 correlative factors upon it in the original body was quite different 

 from that upon the anterior regions. But when we cut the 

 body in such a manner that this level lies at or near the ante- 

 rior end, then the cells adjoining the wound react to the altered 

 conditions by loss of specification and growth, but the final re- 

 sult is dependent in part upon the more posterior levels with 

 which these cells are correlated. The question as to whether 

 this correlation determines the character or quality of the reac- 

 tion or merely its rate or intensity will be discussed in connec- 



