dredged up from the Gulf of Manaar. 455 



developed from the cccnosarc extending over foreign sub- 

 stances enlarged, dichotomously branched one or more times, 

 and thus rendered dendriform (fig. 8, «, g). Largest and 

 most perfect specimen, which is that figured, about 2 inches 

 long and If inch in its broadest part ; stem of the same at 

 the lower end, which has been broken off, l-60th inch in 

 diameter. 



Hob. Marine, growing on hard objects. 



Loc. Gulf of Manaar, in Q5 fathoms, N. lat. 7° 18'. Western 

 coast of Ceylon. 



Obs. It may seem premature to some to establish a new 

 j-enus of Hydractiniidse on the mere skeleton of a Hydroid 

 Zoophyte ; but the characters of the family are so peculiar 

 that, to those acquainted with the species which have been 

 described, there is little risk of error in this respect. The 

 generic distinction, however, is founded on the solidity of the 

 structure, which in all other instances hitherto noticed is 

 formed of reticulated fibre, like that of Hydractinia ecliinata. 

 From the number of fragments of Hydradendrium spinosum 

 among the Manaar dredgings, it would appear to be very 

 plentiful in this locality. Under the microscope, when fully 

 divested of the coenosarc, the regularity of the spination and 

 clear amber-colour of the whole stem, make it a beautiful 

 object. In development, of course, the skeleton is formed 

 upon the embryonic sarcode, which thus becomes separated 

 from the coenosarc of the exterior, although it is probable that 

 subsequently the latter, which must have sprung from the 

 former too, in the first instance, adds the greater portion of 

 the horny material in layers, like the sarcode of the sponges 

 to their structures both horny and siliceous. That the coeno- 

 sarc can do this is proved by the forest of dendriform spines 

 which it develops from its surface when extending from the 

 stem to other bodies, as shown in the illustrations (fig. 8, a &g) . 



EXPLANATION OF THE PLATES. 

 Plate XVIII. 



Fig. 1. Polytrema cylindrieum, n. sp., in situ, showing the sponge-spieules 

 pi'ojecting from the apertures in the ends of the branches respec- 

 tively : a, horizontal section, to show the cortical and medullary 

 cell-structure (magnified 2 diameters) ; b, form of cell about the 

 ends of the branches, broken open ; c, aperture with sponge- 

 spieules on their way to the interior ; d, the same in the inte- 

 rior ; e, portion of surface of pillar, to show the arrangement of 

 pores &c. ; /, fragment of outer wall of cortical cell, cut hori- 

 zontally to show the sunken pore-tubules ; g q, tubules. (All 

 on scale of l-48th to l-1800th inch.) 



Fig. 2. Tubipora reptam, n. sp., in situ, showing the calycles and their 



31* 



