STRUCTURE AND DEVELOPMENT OF COELOl'LANA. 29 



the outer ring are rather short and stiff and project freely into the Uimen 

 of the canal. As Chun ('80) has remarked, the cilia appear to be 

 cemented together into a mass somewhat reminding one of a comb of 

 the rib of ordinary ctenophores. 



The rosette is found practically in any part of the ciliated endoderm- 

 al epithelium of the canal-system, except possibly that of the epithelium 

 of the infundibulum and of the perradial canal. Its situation is frequently 

 marked by a funnel-shaped depression of the canal wall. 



As to the function of the rosette, most of the previous writers 

 have regarded the structure to be a passage through which the water 

 in the canal flows into the body parenchyme. Judging from its appear- 

 ance in general, but especially from the condition of the cilia, this is 

 very likely to be the case. 



As already mentioned, the vacuolated epithelium encloses food 

 particles in the cell-body, of which pieces of striped muscles apparently 

 of small crustaceans are the commonest. This fact indicates that, there 

 takes place intracellular digestion in the endoderm, just as in ordinary 

 ctenophores. 



Besides food particles, the cells may contain an abundance of 

 detrita, either organic or inorganic, which may be found to be enclosed 

 in vacuoles. The part of the cell-body surrounding such a vacuole or 

 vacuoles often comes out from the cell and forms a more or less 

 distinct spheroidal body (PI. 4, fig. 15, PI. 5, fig. 4; rw). Such a body 

 is eventually discharged or extruded, as it were, from the seat of their 

 origin into the canal, to become the free corpuscles that float in, and 

 move along with the water in the canal system. The corpuscles may 

 be observed in various stages of extrusion on sections. On completion 

 of this process, there are left in the matrix hollow spaces which nearly 

 correspond in size with the liberated corpuscles. After repeated occur- 

 rence of the process, the matrix may present a much pitted or honey- 

 combed appearance. 



The corpuscles set free from the digestive epithelium are conveyed 

 through the canal system evidently by the action of the cilia carried by 

 the low epithelium'. In living specimens, the circulation of the corpuscles 

 offers a very remarkable spectacle under the microscope, as was noted 

 by Abbott ('07). It presents an appearance somewhat reminding us of 

 the circualtion of blood- corpuscles through the blood-vessel. The direction 

 of the current may differ within a short distance and may be reversed 

 in one and the same place within quite a short duration of time. The 



