NATURAL HISTORY OF THE SHIP-WORMS. 213 
the afferent branchial vein, and of unknown function. These three glands described 
by Deshayes are parts of one and the same structure, which is present, in different 
degrees of development, in all of the three species I have studied. In all three the 
part in the gill is well developed. In 7. dilatata the umbonal portion of the gland 
is so large as to occupy a considerable part of the umbonal region of the shell cavity; 
in XY. gouldi it is small, and in 7. navalis apparently rudimentary. In his studies 
of the pericardial glands in lamellibranchs, Grobben sought in Teredo for the gland 
described by Deshayes in the umbonal region, thinking it might represent a part of 
the pericardial gland of other forms which possess this organ. He failed to find it 
and supposed it to be absent. However, though he apparently had none of the 
forms with which I have worked, J think it was doubtless present in his species. 
In the larva this peculiar structure is present on each side in front of the 
cerebral ganglion, though still comparatively simple (g D, fig. 7). It is vesicular 
and filled with spherical cells of apparently mucous nature. A duct leads to the 
exterior, opening at the side of the mouth on the ventral side of the velum. 
The structure of subsequent stages of the gland will be best understood by 
first describing that part in the gill. An examination of figure 38 will show that 
this modified portion contains elements of two very different types of structure. 
Their distribution and relations are best shown in figure 37, which represents a 
lamina from the gill of 7. navalis, in which species they are most strikingly developed. 
This figure also shows the distribution of the gland in the branchial vein, and that 
this portion is of the same nature as that lying in the lamina adjacent to it. Still 
farther from the vein is the second type of structure. Ramifying in all directions 
from the latter are dentritic processes which penetrate the epithelial walls of the 
lamina. These ramified portions are the primary structures, apparently, and the 
other two are derived from them. The structure of the dentritic portion is shown in 
figures 46 and 47, which were drawn under a magnification of 1,900 diameters. The 
processes seem devoid of any membrane. The contents consist of very minute 
filamentous structures arranged lengthwise in the direction of the process. Lying 
in the mass thus formed are nuclei which vary in number and position. The mid- 
dle one in figure 48 indicates that they may change position, and that the whole 
structure forms a syncytium. The enlarged portions of the processes, shown in 
figure 37, become surrounded by «a specialized epithelial covering, apparently 
derived from the Jining cells of the gill lamina. This stage is represented in figures 
46 and 47. The minute filamentous structures have taken on a more irregular 
arrangement, and lying within the mass are spherical cells of varying appearance. 
While some (fig. 45) are coarsely granular, others are almost homogeneous. The 
nuclei Jie on one side of the cells. 
The other type of structure (fig. 49-51) I am confident, though not perfectly 
sure, is also derived from the dentritic processes along with a modification of the 
surrounding epithelium. The developed structure is of remarkable appearance 
(fig. 51). The base is composed of modified epithelium cells of the wall of the 
lamina. The nuclei stain lightly and lie in a granular protoplasm, from which 
deeply staining rods project toward the blood space of the gill Jamina, but from 
which they are separated by a membrane formed of very flat cells. The develop- 
ment of this structure seems to be as follows: When the dentritic processes penetrate 
