286 University of California Publications. [Zoolog-j 



Contraction of the stem does not appear to be the only method 

 of increasing its diameter and producing a tension on the mem- 

 brane. It is noticeable in the figures in Fig. 29 that there is an 

 upward arching of the partition. The latter at this stage is an 

 unbroken sheet of cells completely separating the cavities of 

 hydranth and stem. The cavity of the stem, then, has no outlet. 

 Under this condition, a rise in osmotic pressure of the liquid 

 filling it would distend its walls wherever possible. The non- 

 muscular parenchymatous partition would be forced outward 

 and its cells put upon the stretch. The degree of outward dis- 

 placement would depend partly upon the pressure in the hy- 

 dranth cavity. The tension on the membrane, however, would 

 not be affected by that fact, provided the walls of the stem were 

 not rigid, non-extensible at the level of the partition ; and they 

 are not rigid, as we have seen. 



In the larval stages of development, the same pressure may 

 be a considerable factor in the elongation of the stem. There 

 is evidence that it exists even after the peripheral canals have 

 become well established ; but at that time it can have little effect 

 upon the form of the stem in comparison with the turgid axial 

 endoderm to be considered below. 



(5) It is hardly to be questioned that the cells in the mem- 

 brane, once it be established, add to their numbers just as the 

 other structures in the body do, by proliferation. There is no 

 evidence, however, that cell division plays an active part in the 

 extension of the membrane. 



(6) The fenestrae develop irregularly between the cells of tht 

 membrane in connection with the peripheral canals. They are 

 not present from the beginning, the membrane being unbroken 

 even when the polyp possesses nine proximal tentacles, a mouth 

 and a well differentiated digestive epthelium in the hydranth. 

 It is suggestive that they appear at the points where the corre- 

 sponding peripheral canals, with outer walls of ciliated cells, 

 could deliver columns of water against the membrane. 



(&) The development of the peripheral canals may now be 

 traced. Fig. 32 represents a section from the middle stem re- 

 gion of the same polyp whose fenestrated membrane is shown in 



