280 Mr. H. M. Bernard on the 
When the corallum grows in thickness and the basal 
portions die, a secondary epitheca appears under the edges of 
the living layer as it creeps over the older dead portions. In 
Turbinaria the earliest buds arise from the parent polyp 
above and free from the epitheca, which merely spreads out 
as a cementing substance beneath the expanding base or foot 
of the cup-shaped corallum. But here again also a secondary 
epitheca, which can have no connexion with the original 
epitheca, may develop under flat horizontally expanding 
leaves and under the edges of incrusting glomerate forms. 
The two genera, however, rise up so very differently from their 
respective epithece that there are no grounds for attempting 
to deduce one from the other. 
(6) The ccenenchyma in Astrwopora is far more highly 
specialized than it is in Turbinaria. 
(c) In Turbinaria the intracalicular portions of the septa 
are, as arule, highly developed, uniting at the base of the 
fossa to form the conspicuous columella. The extracalicular 
portions show no special development, but merely form the 
ccenenchyma connecting the polyps. In Astrewopora the 
intracalicular portions of the septa are very feebly developed, 
only the first cycle even approaching the centre. As each 
polyp advances in age and length, the new cycles, second and 
third, are less and less developed. On the other hand, the 
extracalicular or costal portions of the septa are highly 
specialized, forming the characteristic ccenenchyma of the 
enus. 
While both genera therefore are coenenchymatous corals, 7. e. 
in both the whole corallum which rises freely above the epitheca 
is a product of the septa, and may be related, we have no 
evidence to support such relationship ; indeed, they differ from 
one another as far almost as such corals could differ, from the 
very outset of their respective life-histories. 
Parasites and Commensal Annelids.—It is worth noting 
that the Balanids which infest some of the highly echinulate 
Astraoporans defend themselves from being overgrown by the 
coenenchyma by formidable series of spikes on their shells, 
which closely resemble in size the echinule of their hosts. 
The Balanids infesting the Turbinarians with their less 
echinulate coenenchyma are detended by a much less pro- 
nounced arrangement of teeth. 
There are several records of commensal Annelids whose 
calcareous tubes are worked up with the skeletal formations 
of corals. The most remarkable case, perhaps, is that of 
