DR. ALLMAN ON THE ANATOMY AND PHYSIOLOGY OF CORDYLOPHORA, 371 
obviously true secreting cells, destined by their own rupture and that of the mother- 
cells to discharge their contents into the cavity of the stomach ; while the free 
granular masses in the interior of the mother-cells were most probably originally 
contained in secondary cells, of whose secreting action they are the product, and 
which at a subsequent period had disappeared. 
{c.) Tentacular canals . — The tentacula are hollow processes from the sides of the 
polypes, and their cavities are simple continuations of that of the stomach. This 
view is certainly opposed to the appearances presented on a superficial examination, 
and is at variance with the accounts given by other observers of the structure of the 
tentacula in the marine Hydrolda ; very careful and repeated observations, however, 
have convinced me of its truth. It must be admitted, that at first sight, the tentacula, 
even under well-managed microscopical examination, have exactly the appearance of 
tubes whose cavity is interrupted at regular intervals by completely formed transverse 
septa; I have, however, satisfied myself that in Cordylophora, and probably also in 
all the other Hydroida, the tube of the tentacula is perfectly continuous. The 
tentacula consist in reality, like all other parts of the animal, of an external or 
ectodermic layer (fig. 9 a, a), which is a simple continuation of the general ectoderm 
of the body, and of an internal or endodermic layer (fig. 9 6, b), which is in the same 
way a continuation of the endoderm of the stomach. The tentacular ectoderm has 
been already described. The endoderm consists of rather large cells with very deli- 
cate walls ; it constitutes a thick lining of the tube, and nearly fills up the entire 
cavity ; and indeed this cavity, by a temporary approximation of its walls, appears 
capable of occasional obliteration. The cells of the tentacular endoderm are glan- 
dular, and may be generally seen to contain the peculiar brown granules character- 
istic of the general endodermic secretion. The appearance of transverse septa is 
probably due to the occurrence at regular intervals, of interruptions in the continuity 
of the endodermic layer. 
Though the tentacula of Hydra present no appearance of the transverse septa, their 
structure is nevertheless essentially the same as that of the tentacula of Cordylophora, 
the only difference being in the fact that the endoderm forms a thinner stratum in 
Hydra, and thus encioaches less on the cavity of the tube, and does not present the 
successive interruptions which in Cordylophora give rise to the fallacious appearance 
of transverse septa. 
{d.) Canal of the Coenosarc. — The cavity of the stomach is continuous posteriorly 
with the canal of the coenosarc, fig. 3 h. The rounded masses of cells constituting 
the large rugae into which the endoderm of the stomach is thrown, cannot be traced 
further back than the posterior extremity of the clavate body of the polype, and from 
this point, which may be conveniently though somewhat arbitrarily assumed as the 
posterior termination of the stomach, the canal of the coenosarc commences, and 
thence extends as a continuous tube through all the ramifications of the coenosarc. 
In that portion of the canal which immediately succeeds to the stomach, and which 
