AND DEVELOPMENT OF PYROSOMA. 
205 
the neural side, the two peripharyngeal ridges pass on to the elevation of the inner tunic 
in which the ciliated sac opens, and unite upon its posterior half, widening as they do so, 
whence their junction forms a triangular area with its apex directed backwards. 
The very singular structure which I formerly termed the endostyle, and which I was at 
one time inclined to regard as a kind of internal shell, is, in reality, a longitudinal Ibid or 
diverticulum of the middle of the haemal wall of the pharynx, which projects as a vertical 
ridge into the haemal sinus, but remains in free communication with the pharynx l>y ;i 
cleft upon its neural side. In consequence of the thickness and opacity of t lie epithelium 
which lines the fundus of this fold, it appears (especially in the fresh state) like a stromr 
hollow rod mounted .upon a thin ridge-like plate. 
Transverse sections, however, demonstrate the true nature of this structure w i 1 1 1 perfect 
clearness (PI. XXX. fig. 8). The bottom of the diverticulum is seen to be occupied by two 
stout cords, formed of elongated epithelium- cells set perpendicularly to the axis of t he cord. 
These cords are separated from one another by a slight interval. Externally and below 
they are in contact with two lateral cords of similar cells. Anteriorly 1 he lateral pass into 
the middle cords, while the latter project beyond the anterior boundary of the groove-like 
entrance into the cavity of the endostyle (and, consequently, of the anterior ends of the 
lips or epipharyngeal folds which bound it) and, coated by a process of the inner tunic, 
constitute the free, rounded, anterior termination of the endostyle. 
Posteriorly, the same confluence of the median and lateral cords takes place; but here 
the endostyle extends much further beyond the limit of the groove and its bounding 
folds, and constitutes a free, hollow cylindroid or conical process, which, as we shall see, 
plays a very important part in the process of gemmation, where I shall have occasion to 
speak of it as the endostylic cone. 
The hypopharyngeal band is not, as in many Ascidians, separated for the greater part 
of its length from the neural wall of the ascidiozooid. On the contrary, in consequent 
of the position of the oesophageal aperture close to the neural wall of the hranchial 
cavity, and the non-extension of the atrium forwards in the middle line, the hypo- 
pharyngeal band is represented only by the inner tunic of this neural wall, which hes 
parallel with the outer tunic, and is separated from it only by the neural sinus, wluch 
usually contains a great aggregation of blood-corpuscles. These corpuscles arc com- 
monly aggregated more densely in the posterior two-thirds of the hypopharynuval sinus, 
and not unfrequently are divided, more or less completely, into two lateral portions by a 
median clear space. When this state of things exists, the hypopharyngeal sinus, under a low 
power, presents exactly that appearance which is figured by Savigny as a s.phon-l.ke tube 
The inner tunic of the hypopharyngeal band is produced in the middle line into eight 
slender conical processes-the lemguet*, which are situated at tolerably equal distances from 
one another. Thus both the neural and the hemal walls of the pharynx are separated 
from the outer tunic in the middle line by nothing but the corresponding sinuses ; and 
the same holds good of the lateral waU in the region of the periphary n geal ridge. But, 
at any point behind this, either a vertical and transverse section, or a view from above, 
shows that the inner tunic (or pharyngeal wall) is separated from the outer unic by a 
more or less wide space, enclosed wit bin a membrane which is totally distinct from both 
