THE STARnSH SOLASTER ENDECA. 29 



origin of pouch VI. The further extension of the canal follows the development of the 

 later-formed pouches, and it is only after the preoral coelom has almost entirely dis- 

 appeared that the tips of the dorsal and the ventral horns of the hydrocoele crescent 

 meet in interradius IX/I, thus completing the ring-canal of the water vascular system. 

 The remains of the stalk of connection of the preoral coelom with the axial sinus are 

 traceable as passing oralwards within the completed ring (p. 27). In preparation for 

 their meeting, both ends of the crescent, but particularly the dorsal one, curve slightly 

 to the aboral (right larval) side. This bending to the right is evident in the dorsal 

 horn during the middle free-swimming stage, but in the veatral horn only at the 

 commencement of metamorphosis. 



It should be stated that for a time pouch I of the hydrocoele does not have the 

 chance of projecting into the posterior coelom. To begin with, this pouch pushes out 

 into the somewhat thickened mesenchyme of the anterior ccElomic wall, to which is added 

 afterwards the tissue produced in this region through the reduction of the preoral 

 coelom. Finally, the ventral horn of the posterior ccelom extends round to the madre- 

 poric interradius, displacing the tissue in question. By this time radial canal I is 

 embedded in the oral wall of ray I, but the ampullae belonging to the tube-feet of this 

 ray can project freely into the ccelomic cavity now provided. 



Dorsal Sac or Madreporic Vesicle. — The origin of the dorsal sac or madreporic vesicle 

 has already been noted (p. 26). Almost immediately after its separation the cells 

 lining it become much flattened. In the larva it lies on the ventral side of the 

 hydroporic canal, and in the young Starfish just after metamorphosis it will be found 

 close under the surface, slightly nearer the centre of the disc than the hydropore and 

 somewhat to the dextral * side of that opening as one looks from the aboral side. This 

 also is the position it occupies in the adult with reference to the axial coelom and the 

 madreporic plate. In early metamorphosis its outer wall, /. e. that adjacent to the 

 axial coelom, becomes invaginated by a fold containing a group of cells continuous 

 with those forming the axial organ within the axial coelom. Later, the tissue thus 

 folded in the dorsal sac acquires a highly vesicular appearance. At the very end of 

 metamorphosis a new element is added to the interior of the fold. This is an ingrowth 

 from the posterior coelom of the kind of cells which Macbride has described in Asterina 

 as giving rise to the genital rachis (see under posterior coelom, p. 36). 



It will be seen that the place of origin of the madreporic vesicle would suit a 

 homology with the hydrocoele. In no instance, however, that I came across did it 

 actually develop a hydrocoele-like structure as in the examples of Asterina put on 

 record by Macbride. Once I found it provided with an external opening on the dorsal 

 side close to the hydropore, but this was in a specimen which showed various other 

 abnormalities. No enautiomorphic specimens were recognised. 



Axial or Central Part of Anterior Coelom. — It remains now to give an account of the 

 axial or central part of the coelom, i. e. that part which remains after the separation of 



* See note on p. 13. 



