TIIK AXIAL GRADIENTS IN HVDKO/OA. 413 



had previously shown a lack of difference in rate of regeneration 

 of middle and basal thirds in T. crocea. Differences in the growth 

 habits of the species are in harmony with the experimental 

 findings. In T. marina (see illustration in Child, '15, p. 90), 

 there are no branches but the base of each stem runs along the 

 substratum as a stolon for some distance, then turns vertically, 

 and gives rise to a new hydranth at its tip. This method of 

 growth indicates that the hydranth in T. marina dominates a 

 considerable length of stem. The Atlantic coast species, on the 

 other hand, branch freely, frequently at a relatively short distance 

 from the terminal hydranth. This fact suggests that the domi- 

 nance of the terminal hydranth does not extend very far proxi- 

 mally. Beyond this limit the axial differences along the stem 

 would be slight or absent. It it -probable that if shorter distal 

 pieces of the stem of T. crocea were cut into thirds, a difference 

 would be found in the regeneration rate of the middle and basal 

 thirds. In both my and Banus' experiments, however, rather 

 long stems were used. 



3. Relation of Rate of Regeneration to the Length of the Piece 

 In a previous paper ('20) I reported that the length of the piece 

 has little effect on the rate of regeneration when diameter and 

 level of the distal cut are constant, except when the pieces are 

 very short. This result was again verified on T. crocea and on the 

 Maine species. Thus in 16 pairs of pieces of T. crocea, the two of 

 each pair being 10 and 5 mm. long, respectively, with the distal cut 

 taken at the same level in both, the 5 mm. pieces regenerated oral 

 hydranths first in eight cases, the 10 mm. first in eight cases. It 

 is understood that the diameter of the apical end was the same in 

 the two members of each pair. At Swan's Island an experiment 

 was performed comparing the rates of regeneration of pieces 

 15, 10 and 5 mm. long, the distal diameter and level of the distal 

 cut being the same in the three pieces of each set. The rate of 

 regeneration of the 5 and 10 mm. pieces was equal throughout 

 this experiment; and this was also the case with most of the 15 

 mm. pieces, but a few of them preceded in production of hydranths 

 the shorter pieces by a short time interval. 



When the pieces are shorter than 5 mm. in length the produc- 

 tion of hydranths is greatly slowed down, so that the regeneration 

 27 



