19J.0J Torrey: Regeneration of Hydranth and Holdfast. 207 



in longitudinal sections, there is a terminal area of the body wall 

 that is much attenuated just after the wound has closed. Il 

 appears to be a proliferating area, definitely bounded by the cul 

 edge of the old mesogloea, its own mesogloea negligibly thin. Its 

 attenuation suggests the formation of the fenestrated memhrane 

 in the larva, of wound closure in Tubularia, and is probably 

 governed partly by similar factors. The new hydranth cavity is 

 formed by the degeneration and disappearance of the axial endo- 

 derm cells at the end of the stem, the tip of the stem becoming the 

 tip of the new proboscis. The withdrawal of the central core of 

 axial endoderm puts the peripheral canals, indirectly, in com- 

 munication, since they are normally bounded by axial endoderm 

 on the inner side (fig. 2) . Between (hem. however, the axial tissue 

 remains for a time in the form of ridges (fig. 3) which are more 

 or less prominent according to the size of the stem. By the dis- 

 integration of the largest, must highly differentiated cells' and a 

 decrease in size and an increase in granulation of others, it 

 is converted finally into epithelium whose cells resemble tin- 

 epithelial cells lining the peripheral canals (fig. 3). The 

 epithelium of the hydranth cavity is thus derived from two 

 sources. At first it is low, scarcely thicker than the ectoderm 

 surrounding it. Gradually its cells become long and the low 

 narrow folds between the canals give way to the much broader 

 folds of what has now become the chief digestive epithelium of 

 the hydroid. 



Even before the wound has closed, in some cases (fig. 6). but 

 usually afterward, the tentacles begin to appear. The proximal 

 tentacles usually arise first, in two sets. The first set (fig. 4) 

 appears as a circlet of conical elevations just above the lower 

 limit of the hydranth cavity, one to a canal. Though the canals, 

 by the obliteration of the axial endoderm, may be parts of the 

 hydranth cavity at this time, they are still separated by the 

 ridges previously mentioned, which are continuous with inter- 

 canalicular tissues below. 



The number of the tentacles in this set usually corresponds 



3 Such loss, in connection with regeneration, of tissue so specialized as 

 apparently to be incapable of a change of function, will be considered with 

 the various phenomena of retrograde metamorphosis in another paper. 



