330 ON THE ANATOMY AND DEVELOPMENT OF PYROSOMA 
tunic (or pharyngeal wall) is separated from the outer tunic by a 
more or less wide space, enclosed within a membrane which is totally 
distinct from both the outer and the inner tunics, except at the atrial 
orifice, where it passes into the former, and at the anus,. where it 
becomes continuous with the latter, 
Except for these two breaks, the membrane in question (which is 
the third tunic of Milne-Edwards, and is what I have elsewhere 
termed the a¢ria/ tunic) might be compared to a closed serous sac, 
reflected over the viscera, on the one hand, and over a part of the 
external tunic, on the other, but leaving a space between itself and 
both these parts, which space is filled with blood, and forms a part 
of the general system of sinuses of the body. A careful examination 
of the side view (Pl. XXX. [Plate 29] fig. 1), the upper view (fig. 3), 
and the transverse section (fig. 2) will render this statement 
intelligible. 
In the first, the atrial tunic is seen to be reflected over the 
posterior face of the stomach and first part of the intestine, and then 
to form the roof and the floor of the cavity, which lies between the 
intestine and the atrial aperture, and which I shall term the mzd- 
atrium. In fig. 3, the atrial tunic is shown to be continued forwards 
at the sides of the intestinal canal on to the pharynx, united with 
which, it forms the branchial sac. Arrived close behind the peri- 
pharyngeal ridge, it is reflected on to the external tunic, and then 
passes directly backwards to the atrial aperture. The testis and 
ovisac, which are seen, in this view, over the alimentary canal, lie 
altogether above the roof of the mid-atrium (fig. 1), and therefore do 
not in any way interfere with the free and wide communication of the 
mid-atrium with the two spaces, or /ateral atria, which lie between 
the branchial sac and the body-wall, and are well shown in the 
transverse and vertical section (fig. 2). Both in this section and in 
fig. 3, short cords are seen to pass between the parietal and the 
visceral layers of the lateral atria. They are hollow, and place the 
parietal sinuses in communication with those of the branchial sac. 
It follows from what has been said that the wall of the branchial 
sac of Pyrosoma (and, I may add, of all Ascidians with a similar 
respiratory apparatus) consists, internally, of that portion of the 
alimentary tract which lies in front of the cesophagus and behind the 
mouth (or, in other words, of the pharynx), and, externally, of the 
visceral layer of the atrial tunic. Now, these two membranes do not 
remain entirely separated by the interposed sinus, but are united at 
regular intervals, so as to give rise to hollow vertical bars separated 
by equally long vertical clefts—the branchial stigmata. 
