552 
PROFESSOR ALLMAN ON THE STRUCTURE 
numerous refringent corpuscles and a few brown granules. No boundary membrane 
was evident in any of these cell-bodies. At the inner or free surface the endoderm of 
the whole of the gastric cavity, except in the region immediately below the mouth, forms 
long conical processes, which project like villi into the cavity (figs. 1 & 2, b). These 
processes, like the more external parts of the endoderm, are mainly composed of large 
cells, formed of clear protoplasm, with nucleus and refringent corpuscles ; but besides 
these there exist also towards the free ends of the processes numerous smaller spherical 
cells (fig. 2, c ), loaded with dark-brown granules. These cells are most abundant in 
the villi-like processes which are developed towards the proximal end of the body. They 
form a much less coherent tissue than the large clearer cells, and may be easily isolated 
under the microscope. Indeed they are constantly being thrown off, and may be often 
seen to be voided through the mouth of the living animal. 
Extending over the free surface of the endoderm is an exceedingly thin stratum of a 
clear homogeneous protoplasm (fig. 2, d). This protoplasmic stratum is most obvious 
the villi-like processes, where it has the property of developing very minute, irregular, 
pseudopodial projections (eee), which are constantly changing their shape, and may 
be seen under the microscope to be slowly protruded and withdrawn. The free surface 
of the endoderm carries also long, very slender vibratile cilia. I believe that the thin 
layer of protoplasm which extends over the free surface of the endoderm is continuous 
with an interstitial undifferentiated protoplasm which exists in small quantity between 
the endodermal cells. Its occurrence, with its pseudopodial extensions, on the gastric 
surface of the animal is full of interest, and suggests a close analogy between the absorptive 
action of the gastric surface and amoeboid reception of nutriment ; more especially when 
we bear in mind that the cells between and over which the semifluid protoplasm is spread 
are destitute of membrane, and that their protoplasm must be in direct relation with that 
of the pseudopodial stratum. 
The cilia are extremely fine and difficult of detection. They do not appear to be con- 
tinuous over the whole gastric surface, but to exist only at intervals. They probably 
originate directly from the proper surface of the endodermal cells, in which case they 
must traverse the pseudopodial layer. They may, however, be direct processes of this 
layer. Indeed it is difficult in either case not to regard them as modified pseudopodia. 
True vibratile cilia, like pseudopodia, can originate only from the surface of membrane- 
less protoplasm, which thus possesses, as one of its characteristic properties, the faculty 
of being able to develop two kinds of processes — the non-mutable vibratile cilium and 
the mutable pseudopodium. 
From the gastric cavity the endoderm is continued in an altered form into the cavity 
of the tentacles (fig. 2, b, & 3). Here its condition differs strikingly from that of the ten- 
tacular endoderm of other marine hydroid trophosomes ; for instead of forming the clear 
septate core which is so very characteristic of these, it consists of a single layer of small 
round cells surrounding an open axile cavity, and so loaded with opaque granules that 
the axis of the extended tentacle appears nearly white under reflected light. 
