ON THE ANATOMY OF FISHES. 
211 
abnormal forms as Loricaria Callichthys and Hypophthahnus constitute the ventral 
walls of the bony capsules for the air-bladder. There is no trace of any superficial 
fibrous investment to the bladder, or of a transverse membrane, in fact, the place of 
these structures is really taken by the ventral outgrowths which here, as in other 
instances, may, with some probability, be regarded as due to the ossification of 
the former. 
The tripus is very similar to that of Glyptosternum, the crescentic process having so 
slight an inward curvature as to be almost straight. The intercalarium is a very small 
nodule of bone imbedded in the interossicular ligament. Except that it has no 
ascending process, the scaphium is normal. 
The cavum sinus imparis and the atrial cavities differ in no essential respect from 
those of the more normal Siluridae, but, unfortunately the auditory organ in the only 
example of this species that we were able to examine was too badly preserved to 
admit of any reliable results as to its structure being obtained. 
Schilbichthys garua. 
In its skeletal modifications this species bears some resemblance to Cryptopterus 
microims, and also, but in a lesser degree, to Bagarius, Glyptosternum, and other 
similarly modified Siluroids. As we have already pointed out, the agreement between 
Schilbichthys and Eutropiichthys is very striking. The air-bladder is said by Day (9) 
to be “ small, somewhat heart-shaped, and closely attached to the bodies of the 
anterior vertebrae” {loc, cit., p. 709). 
The first, the complex, and the fifth and sixth vertebrae are rigidly connected with 
one another and to the skull. The body of the first vertebra is very small, and 
inferiorly appears to be completely enclosed within the contiguous concavities of the 
basioccipital and complex centrum. The length of the complex centrum (fig. S9, c.c.) 
is nearly twice that of the fifth centrum {v}), and the latter, in turn, bears about the 
same proportion to the normal vertebral centra which succeed it. The neural arches 
of the complex and two folio wdng vertebrae are confluent, or, at the most, are only 
partially separated by faint irregular sutures. The spine of the third vertebra is 
represented by a thin ridge of bone which thickens anteriorly, and slopes upwards and 
forwards to its articulation with the supraoccipital and exoccipital ; posteriorly, the 
ridge is continuous with the base of the long and stout spine of the fourth vertebra, 
which inclines obliquely backwards, overlapping and almost suppressing the spinous 
process of the fifth and sixth vertebras. The seventh vertebra {vJ) is quite free. 
The transverse process of the fourth vertebra {t-p-^) has a broad root, slightly 
concave on the ventral surface, and a distal half, which is greatly expanded, and has 
its anterior and posterior margins curved directly downwards in such a way that they 
form the sides of a short but relatively deep, transversely disposed, bony semi-cylinder, 
the cavity of which is open distally as well as ventrally in the dry skeleton. In the 
