ON THE ANATOMY OF FISHES. 
181 
becomes a deep groove, which is open in front, at all events so far as the transverse 
process itself is concerned, and gradually dies away towards the distal end of the 
process. The groove is not an open one, inasmuch as its superior margin joins the 
inferior edge of the “ os occipitale externum ” (post-temporal) and its inferior margin 
the posterior edge of the same bone, so that the groove becomes a closed canal, except 
that it has a distal aperture between the free edge of the transverse process and the 
overlapping '‘os occipitale externum.” Near the distal aperture of the canal there is 
a small area of soft skin, between the cranial plates and those which invest the lateral 
surface of the trunk. 
The cavity enclosed by the proximal portion of the transverse process is separated 
from the outer groove by a slight ridge, and contains a globose membranous bladder, 
which, from the analogy of Acanthicus, Reissner rightly regards as equivalent to one- 
half of the air-bladder. From the inner wall of the cavity of the transverse process a 
narrow canal extends inwards and forwards, and in this canal lies the almost rectangu- 
larly bent tripus.'" This ossicle is said to consist of two processes united at a right 
angle, and of these one is transversely disposed and connected with the wall of the cor- 
responding air-sac, while the other, or anterior process, projects from the canal in which 
it is lodged into the cranial cavity, a scaphium similar to that of Acanthicus being- 
attached to its free extremity. A membranous canal is said to leave each air-sac and 
to pass through the hinder wall of the cavity of the transverse process, probably for 
the purpose of effecting a communication with the sac of the opposite side. A perfect 
“processus bij ugus ” could not be found, but two osseous processes on the inferior 
surface of the “ first ” vertebra, and a slender, bony spicule which is directed inwards 
and backwards from near the root of each transverse process may possibly represent 
this structure. Possibly, Reissner remarks, the spicules, or tendinous prolongations 
therefrom, may be continuous with the two bony processes. 
No reference is made to the internal ear or to the cavum sinus imparis and its atrial 
cavities. 
We are able to supplement Reissner’s somewhat imperfect descri})tion with a few 
additional particulars. 
The “second” vertebra of Reissner is probably the fifth, and hence, his “first” 
vertebra really represents the complex vertebra of other Siluroids (see Rams.w 
Wright, 44, p. 109). The centrum of the true first vertebra is either wanting, or has 
fused with the complex, but in the latter case no intervertebral suture could be 
detected. The posterior face of the complex centrum has a deep tubular concavity, 
but the almo.st flat anterior face articulates with the contiguous surface of the basi- 
occipital by an interdigitating suture. Almost complete fusion, however, has taken place 
between the arch and spine of the complex vertebra and the supra- and exoccipitals, 
but from the position of the external atrial apertures, and the mode of articulation of 
* This canal is really a simple aperture between the exoccipital and the arch of the comple.x vertebra, 
and corresponds to that portion of tlio wall of the neural canal which in other Siluroids is fihroiis. 
