296 HYDROID^. Part IV. 



thirty-five of these partitions. The general outline of the cells is cylindrical 

 {Fig. 14''), usually circular, but in certain cases prismatic. From the foregoing it 

 will be seen that the hydra? have no lateral communication with each other, 

 through the mass of the corallum, but that their relations are altogether superficial. 

 A longitudinal section of the cells would seem to show that this is not so, when 

 we find two cells {Fig. 14'', h e) uniting below in one chamber {ni) ; but we have 

 found that this was only the case when the hydrae were down at that level, and 

 consequently superficially related, whereas, at later periods, they were not only 

 separated from the lower chambers by the transverse partitions, but, by the same 

 means, from each other. 



SECTION III 



SERIATOPORA SUBULATA LMK. 



The intervals between the cells at the tip of the branches (PI. XY. Fig. 15, 

 a b c) are as distinctly marked out as in Pocillopora damicornis, and the calcareous 

 deposit equally solid. The borders of the cells at this point are fringed by 

 rather blunt spinules {h), arranged in an irregular row. At the very earliest 

 stages of growth recognizable on the corallum, the young cell possesses a columellar 

 projection, such as is so prominent in the older cells {Fig. 15, j k). Originally, 

 then, these young cells have the form of inverted, truncated cones, which finally 

 deepen and become parallel sided {Fig. lb"), but as they do this the central 

 columella rises, and at the same time, usually, four perpendicular partitions, at ninety 

 degrees from each other, are thrown out from the axis to the periphery, in such 

 a way as to produce four cavities {Figs. 15, / i, and 15", d e) around the axis. 

 After the cell has attained a depth usually equal to its breadth, a transverse 

 diaphragmic partition {Fig. 15% / g) is developed, and then another chamber, or 

 rather a fourfold cavity is formed, to be eventually partitioned off like the pre- 

 ceding one, and so on until the end of the existence of the hydra. Throughout 

 the whole corallum, we find the calcareous deposit solid and amorphous, so that 

 it is not possible that the hydra? should have any lateral communication with each 

 other, excepting at the surface of the colony. At the oldest part of the corallum 

 the spinules are scattered, and have no trace of the serial arrangement which 

 obtains in the younger parts of the branches. 



From the peculiar characteristics of this genus, I infer that the Corallaria Rugosa 

 of Milne-Edwards are more likely to have been Hydroids than true Polyps. 



