26 THE VOYAGE OF H.M.S. CHALLENGER. 
occurs only through the action of reagents. It seems probable that it does occur in the 
living animal, since by its means the masses of large thread-cells are brought as a 
protection directly between the zooid and the exterior. Just as in one small portion of 
the coral the zooids died in the expanded condition, so more often, in certain specimens, 
they die and are preserved with the superficial ectodermal layer not closed in over the 
mouth of the calicle, but with the calicle open, and their retracted tentacles remaining 
fully exposed to view from above. In Plate XIV. figs. 2 and 3, both zooids figured are 
shown in this latter condition. The connection between the superficial ectodermal layer 
within the calicle and the adjacent vascular network of the ccenosare was not made out. 
The layer is probably merely the largely developed ectodermal layer of that part of 
the network, but the connection not having been seen is not indicated in Plate 
XIV. fig. 2. 
The superficial. layer of the ccenosare being a special development of the ectodermal 
cells of the vascular network, and the interspaces in this network being occupied by cal- 
careous trabecule, it follows either that the tips of the trabecule at the surface of the 
ecenosarc must be directly exposed, or that the superficial ectodermal cells of the network 
must close in over them. The latter arrangement seems to occur; and in vertical sections 
of the decalcified coenosarc numerous spaces left by removed calcareous structures are 
seen in the superficial ectodermal layer (see Pl. XIV. figs. 2 and 8), with the ectodermal 
cells arching over to cover them. I should have had no doubt in this matter had I 
not observed that in the living Millepora the soft parts of the ccenosare appear to 
be retracted below the surface of the ccenosteum when the zooids are in their retracted 
condition. It can, however, hardly be the case that any part of the ccenosteum is 
directly exposed to the water. It is probably always covered everywhere by the super- 
ficial layer of the ectoderm, which, however, is in the recent condition so transparent 
as to escape observation. The calcareous tissue of the coenosteum must obviously be 
deposited by the ectoderm, with which alone it is in contact. It spreads by extension 
of the trabeculae at the surface; and since there it is seen to be often in contact only 
with the cells of the superficial layer, it seems that these cells must have the power of 
producing it. The calcareous network undergoes thickening in the deeper parts of the 
living lamina, as must necessarily be the case, because of the formation of the tabule 
and lines of growth, In these parts no doubt the fusiform nucleated cells of the ectoderm 
are the instruments of the deposition of the carbonate of lime. No special calciferous 
tissue was observed, such as exists in Heliopora caerulea. 
Beneath the layer of ectodermal cells in the vascular canals composing the coenosare 
lies a layer of apparently homogeneous membrane, which appears to form everywhere a 
wall to the vessels and canals. The cavities of the vascular network are lined by, and 
in many places nearly filled with, cellular elements of two kinds—pigmented cells and 
small transparent globules. The pigmented cells (Pl. XIV. fig. 9) closely resemble those 
