STRUCTURE AND RELATIONS OE CERTAIN CORALS. 
119 
that they have walls composed in exactly the same manner as those of the calicles, and 
that they are closed below at intervals in the same way by exactly similar tabulae. Further, 
the soft tissues lining the cavities of the coenenchymal tubes are identical in structure with 
those lining the calicular cavities, and the same transverse system of canals connects the 
summits of the tubes with one another and with the summits of the calicular cavities. 
It seems by no means improbable that the ccenenchym here is composed of the tubes 
of absorbed polyps or zooids which have lost the rudimentary organs, which they still 
possess in such a form as S'arcophyton , and have become mere tubular cavitief, whose 
openings to the exterior even have been obliterated ; it seems impossible otherwise to 
account for the presence of the successions of tabulae in the coenenchymal tubes. The 
foregoing considerations are suggested by the circumstance that a series of fossil corals, 
grouped by M.-Edwards under the Tabulata, appear most probably to have been 
Alcyonarians as well as Heliopora. 
The genus Chcetetes was considered by Keyserling to have belonged to the Alcyo- 
narians, because of the absence of septa in it, and the mode in which its polyps are 
grouped ; but Milne-Edwards retains it amongst the Zoantharians, because of its close 
resemblance to the Favositidae, in which the presence of septa is regarded as conclusive 
in deciding against Alcyonarian affinity. The presence of calcareous septa, however, 
must now be considered a character of less importance than it formerly was. As is 
seen in the case of Heliopora pseudo-septa may exist, which do not necessarily correspond 
in any way, in disposition or number, with the membranous mesenteries. In Stylaster 
and Cryptohelia the calcareous septa are obviously formed as infoldings of the margin 
of the calicles. Here the septa are between, instead of opposite to the tentacles ; and 
membranous mesenteries appear to be absent, or at all events rudimentary only. In 
the Favositidae the septa seem to have been no more perfect than they are in Helio'pora , 
and to have been most variable in number, but often twelve, as also in Heliopora. 
M.-Edwards describes from 10 to 12 septa in Favosites gothlandica. In Michelinia favosa 
30 to 40 subequal septal striae are to be made out at the upper margin of the wall of 
the calicle. I cannot refer to specimens ; but it seems not unlikely that the septa in 
the Favositidae were pseudo-septa as in Heliopora , and that these coralla were formed 
by Alcyonarians, the perforations in the walls having transmitted transverse canals like 
those of Helio'pora and Sarcophyton, and the coralla being free of tabular coenenchym, 
because none of the polyps were aborted as in Heliopora. Some Favositidae seem to 
have formed a compound colony, consisting of polyps and zooids, as Favosites Forbesii , 
where a few large cells are seen set amongst numerous surrounding small ones. Helio- 
lites seems to a certain extent to form a transition stage between a condition such as 
that in Favosites Forbesii and the condition in Heliopora ; for in Heliolites , the more 
ancient form, the coenenchymal tubes are regularly hexagonal, and apparently much 
more nearly equal in breadth to the calicles than in Heliopora. In the growing points 
of Heliopora the hard parts are made up of a series of open, often hexagonal tubes, and 
resemble Favosites in their surface aspect. In Heliopora the transverse canals pass over 
