64 THE VOYAGE OF H M.S. CHALLENGER. 



They appear to be homologous with the radial canals already described as occurring in 

 a similar connection in Sporadopora dichotoma (PL X. fig. 3). In transverse sections of 

 a zooid system, these radial offsets have much the appearance of the mesenteries of an 

 Anthozoan coral cut across, and in some sections they show a certain amount of regularity 

 in disposition at the particular level selected for the cut (PI. XL fig. 12). 



These radial structures are here termed offsets, and not canals, because, although in 

 some instances they appear to be similar in construction to the ordinary coenosarcal 

 canals, and usually show similar structure to these at their outer extremities, they 

 usually consist towards their middles and inner extremities of simple bands of trans- 

 parent fibrous tissue. The exact structure could not be determined, but it seems 

 probable that these radial bands represent radially disposed offsets of the canal mesh- 

 work, which have become developed into fibrous organs with an elastic or muscular 

 function, which is brought to bear on the zooid sac. In the case of the radial canals 

 of Sporadopora dichotoma, distinct muscular elements were observed as forming part 

 of their structure. 



Dactylozooids. — The dactylozooids and their sacs in Allopora profunda are so 

 closely similar in form and structure to those of Stylaster densicaulis, that they need no 

 further description. 



Gastr.ozooids. — The sacs of the gastrozooids in the present species differ from those 

 in Stylaster densicaulis in being of smaller diameter in proportion to the dimensions of 

 the pore cavities, and in being held in place by means of the radial offsets already 

 described. It is possible that the wall of the sac of the gastrozooid lies nearer to the 

 wall of the gastropore in the recent condition of the coral than is represented to be 

 the case in Plate XXXIX. GZ, and in Plate XI. fig. 12; but all the spirit specimens 

 examined yielded a similar result when decalcified. The sac of the gastrozooid is, as 

 usual, a reflection of the surface layer of the ectoderm. In the contracted condition it 

 forms a long tubular cavity, somewhat narrowed in the region just above the gastrozooid 

 and at the mouth, and dilated towards the centre. At its mouth, the marginal fold of 

 the sac rises in the form of a flattened dome somewhat above the level at which the 

 openings of the sacs of the dactylozooids commence. 



The gastrozooids are very deeply seated at the bottoms of their sacs. They are 

 dome-like in form, with expanded bases. A whorl of tentacles, set on at some distance 

 from the summit of the dome, marks the commencement of the large rounded hypostome. 

 The tentacles are twelve in number, and are set on in a single whorl. They are 

 elongate-ovoid in form. From the base of the zooid a series of radially disposed large 

 canals pass outwards to be distributed as in Stylaster densicaidis, a certain number of 

 their branches forming a tortuous meshwork, offsets from which pass to join those of 

 the gastrozooid of the adjacent system. 



The structure of the zooid cyclo-systems is clearly displayed in Plate XL fig. 12, which 



