11 GNIDARIA— G ASTRO -G ANAL SYSTEM 83 



the gastral filaments, in Corals by the tube which leads into the 

 gastric cavity, and in the Ctenophora by the beginning of the gastro-canal 

 system, hitherto unsuitably named the stomach. The mouth of the 

 Hyclromedusm thus leads direct into the endodermal gastric cavity ; 

 whereas the mouth of the Scyphomedusce, Anfliozoa, and Ctenophora leads 

 first into the ectodermal oesophagus, and from that through the enteric 

 aperture, which corresponds with the mouth of the Hydromedusoi, into 

 the endodermal gastro-canal system. 



In all animals, from the Co3lejite7-ata upwards, there is an ectodermal oesophagus. 



The endodermal gastric cavity is met with in its simplest form in 

 the Hydroida as a simple pouch adhering closely to the inner surface 

 of the ectoderm, from which hollow or solid processes extend into the 

 axes of the tentacles. In the long, or principal axis of the body, 

 thickenings of the gastric wall, gastric ridges, which are occasionally 

 4 in number, regularly arranged crosswise, are commonly found. 



We have already described the form^ of the gastric cavity of the 

 Scjiphula in the introduction ; and we also there described the general 

 arrangement of the gastro-canal system in the Antlwzoa. We add 

 here that the septa which separate the gastric pouches from each 

 other round the oesophagus are occasionally broken through near the 

 oral disc by an aperture, so that a sort of circumferential canal arises. 

 The hollow processes of the gastro-canal system in the tentacle axes 

 sometimes penetrate to the exterior at the tips of the tentacles through 

 pores. The free edges of the septa, which are turned towards the 

 principal axis of the Coral individual, are thickened, and are prolonged 

 as mesenterial filaments, freely projecting into the gastric cavity. 

 The epithelium of these filaments contains numerous gland cells, and 

 sometimes stinging cells as well. Some of these filaments, the so- 

 called aeontia, are particularly long and vermiform, and can be 

 quickly shot out of the body, either through the mouth, or through 

 special pores in the body wall. These are found in the Actinia. 



How the gastro-canal system of the Craspedote Medusce (Hydromedusce) 

 can be traced back to the gastric cavity of the Hydrula form, and the 

 gastro-canal system of the Acraspeda {Scypliomedusce) to the gastric 

 cavity of the Scyphula form, has already been shown in the general 

 review. 



The radial canals in the Craspedote Medusce are comparatively 

 seldom limited to the number 4. In most forms their number is 

 greater, and in a few {JEqiioridce) they are very numerous (over 100) 

 and run radially from the central stomach to the margin of the disc, 

 and here enter the circumferential canal. There are also in a few 

 Craspedota, near and between the radial canals, centripetal canals, 

 which start from the circumferential canal and run a certain distance 

 towards, but do not reach, the central stomach. The radial canals 

 may be branched, and these branches may either end blindly or enter 

 the circumferential canal. The central stomach may be divided, the 



