194 
PROFESSORS T. W. BRIDGE AND A. C. HADDON 
curves slightly outwards into the dorsal wall of the contained air-sac as a thin 
tapering process of bone, and at its slender, posterior, or outer extremity becomes 
continuous with the apex of an equally thin, transversely elongated and somewhat 
triangular lamina, the broad base of which is coincident with, and parallel to, the 
attachment of the dorsal margin of the posterior wall of the air-sac to the confluent 
portion of the dorsal lamina (flg. 81). The inner portion of the triangular lamina is a 
trifle longer than the outer, and tapers to a point near the radial nodule, but never- 
theless forms with the root of the crescentic process a characteristic concave, or rather 
angulated inner margin for the attachment of the radial flbres {r.f.) passing therefrom 
to the radial nodule {r.n.). The remaining portion of the lamina is directed outwards, 
like the heel-like process of Macrones, and also tapers to a point. The triangular 
lamina is the only portion of the tripus which is seen imbedded in the dorsal wall of 
the air-sac when the cavity of the latter is laid open from the ventral surface and the 
tunica interna removed. We have several times found the lamina quite distinct 
from the rest of the tripus, but whether normally so, or as the result of accidental 
fracture, we are unable to say. The slenderness of its connection with the tripus, 
and the extremely brittle nature of the bone itself, renders it particularly liable to 
separation by fracture. On the other hand, we have no doubt whatever that in 
many of our specimens the lamina was really continuous with the tripus, and that it 
represents the inwardly curved posterior section of a normal crescentic process. 
The relations and attachments of the fibres forming the walls of each air-sac to the 
tripus are certainly more normal than we expected to find them. As previously 
stated, the inner or mesial portion of the anterior wall of each sac is attached by its 
dorsal edge not only to the radial nodule, but also to the ventral surface of the tripus, 
that is to say, to the root of its crescentic process, although the latter has no trace of 
the ventral ridge which usually marks the insertion of these fibres in the normal 
Siluridae. From this point outwards the fibres of the inner stratum of the tunica 
externa (fig. 81, in.st.) in the anterior, antero-lateral, and lateral or outer walls, 
converge in the dorsal wall towards the tripus, and eventually are inserted into the 
outer margin of the crescentic process from its root to the pointed extremity of its 
heel-like process. The fibres of the outer stratum of the tunica externa have an 
almost exactly similar distribution and relation to the tripus, but pursue the usual 
curvilinear course in passing from their insertion into the crescentic process to the 
antero-lateral and lateral regions of the air-sac. The straight posterior margin of the 
crescentic process appears to be free from any special fibrous connection with the 
walls of the sac, at any rate it is certain that it does not receive the insertion of any 
of the converging fibres of the dorsal wall, and in this respect the tripus of Clarias is 
somewhat singular. The reason for this may possibly lie in the fact that the 
posterior margin of the process is almost coincident with the skeletal attachment of 
the dorsal edge of the posterior wall of the air-sac to the dorsal lamina. 
With regard to the remaining Weberian ossicles we could detect no trace of an 
