ON THE ANATOMY AND DEVELOPMENT OF PYROSOMA = 345 
even in such buds as that represented in fig. 24. The ganglion is 
discernible in buds jth of an inch long (fig. 20) as an opaque oval 
mass situated between the peduncle and the oral end of the bud, and 
very much larger in proportion to the rest of the organism than 
afterwards. The ciliated sac appears as a short cecal diverticulum of 
the pharyngeal cavity, connected with the anterior and hemal side of 
the ganglion. 
A most curious structure is visible in buds 2,th of an inch long, 
and remains obvious until they have attained a length of th of an 
inch or thereabouts. For want of a better name, I will term this the 
‘diapharyngeal band.’ In the section, fig. 21, its upper part is visible, 
passing obliquely downwards and backwards from between the two 
middle bands of the endostyle; while, in figs. 22 & 23, its lower 
extremity is seen to end in the pharynx, immediately over the 
posterior moiety of the ganglion. The diapharyngeal band is hollow, 
and effects a communication between the hamal and neural sinuses ; 
and if, as is possible, the heart of the bud has at this period but 
little functional activity, the existence of this direct channel may 
facilitate the circulation of the blood. However this may be, this 
structure becomes longer and thinner as the development of the bud 
advances ; and all that remains of it, in buds ;4-th of an inch long 
(fig. 24), is a small tubercle which lies over the posterior part of the 
ganglion. Eventually even this disappears. 
I have already spoken of the origin of the branchial stigmata. 
Fig. 21 represents an accidental, but very fortunate, longitudinal 
section of a bud ~,th of an inch long ; the razor having passed rather 
to the right of the middle line above, rather to the left below. As it 
is seen from the right side, the inner surface of the left wall of the 
branchial sac is exposed to view. Of the eleven stigmata, those in 
the middle are the longest and most oval, those at the two ends of the 
series shortest and most rounded. They look clear in the centre, but 
on careful examination they are seen to be closed, the sheet of 
indifferent tissue which forms the innermost wall of the pharynx being 
continued over them. I am strongly inclined to think that it is this 
sheet of indifferent tissue which gives rise to the longitudinal branchial 
bars, for in more advanced buds (fig. 22), in which the median stig- 
mata have undergone much elongation, the same layer is continued 
over their hamal and neural ends, while it has disappeared in the 
interval, except along three longitudinal lines, where it evidently 
forms the foundation of as many longitudinal branchial bars. In the 
more advanced stages, new stigmata are added to the anterior and 
posterior ends of the series. Those already formed elongate, and new 
