54 



gx'owth of the cells of the body wall, accornpanied by mul- 

 tiplication of these s^nvi cells. After some 'weeks the cav- 

 ity of the coelenteron becories dravm out in a diverticul^^m 

 in the direction of the axis of the tentacle, so that the 

 upper part of the gastric cavity becomes stellate in cross- 

 section, Fi?Ture 14 shows this condition in a five months 

 old polyp. This cavity does not reach out into the tenta- 

 cle itself in any of the specimens which I studied, but may 

 do so before metamorphosis takes place. During the whole 

 of larval life, the tentacle is made up of a core of endo- 

 dermal cells in a single row (Fig. 27), as is the case in 

 hydra. Figure 11 shows a polyp with the first pair only de- 

 veloped, and the cell-layers are seen as described. The 

 endoderm cells are filled with a loose protoplasmic net- 

 work (Fig. 27, end), and the nucleus is evident. The con- 

 dition which is seen in an adult tentacle v/ith several 

 cells of endoderm surrounding the central cavity (Fig, 34) 

 is easily derivable from the larval condition, by longitu- 

 dinal fission of the endodermal cells repeated until a 

 cross-section of a tentacle at any level cuts several cells 

 (Fig. 34). 

 B, In the Adu.lt . - The regularity with which the tenta- 



