THE AXIAL GRADIENTS IN HYDROZOA. 89 



were known to have developed within three to four weeks pre- 

 ceding the observations and in which medusa buds or gonozooids 

 had begun to appear only in the basal regions. I have pointed 

 out elsewhere (Child, '196) that medusa-bud development nor- 

 mally begins in colonies which are relatively old physiologically 

 and in the most basal regions of the colony, i.e., in the region of 

 least susceptibility, and as the data of the present paper show, of 

 lowest rate of reduction of permanganate. Later, as the colony 

 grows progressively older, medusa-buds or gonozooids appear 

 farther apically, until finally they may develop even from the 

 apical regions. 



The old colonies with medusa-buds or gonozooids in or near the 

 apical regions may show little or no indication of a general 

 colony gradient. Often in such cases the original hydranths 

 of the most basal levels have disappeared, and in some cases new 

 young hydranths may be developing. Colonies in this stage 

 often show a higher susceptibility and higher rate of reduction 

 in the newly developed, physiologically younger hydranths or 

 growing tips of the more basal regions than in those of the more 

 apical regions, which represent the original generation. Later, 

 when these also have been replaced by a new generation the 

 general colony gradient may once more become like that of the 

 young growing colony. 



In short, the gradient is not a fixed unchangeable condition 

 which can be demonstrated in any hydroid colony. It is a 

 feature of the normal developmental and functional relations 

 of parts, but may be altered, obliterated or even reversed by 

 environmental or physiological conditions. The gradients in the 

 single hydranths and growing tips are more persistent. So far 

 as my observations go, their disappearance is followed by the 

 disappearance through death in the case of more highly spe- 

 cialized, or resorption in the case of less high specialized parts. 

 Similarly, in the colony as a whole the disappearance of the gen- 

 eral gradient is associated with the disappearance, at least 

 temporarily, of the original order and relation characteristic of 

 the colony. Old hydranths or axial complexes may die or be 

 resorbed and new ones may develop, but my observations indicate 



