ON THE ANATOMY OF FISHES. 
227 
more or less completely invested by the modified transverse processes of the fourth 
and fifth vertebroe, but the latter do not form deep concave recesses or capsules for 
the partial or complete enclosure of the entire air-bladder. 
In the Siluridoe abnormales on the contrary, the air-bladder is ahvays very small, 
or even diminutive, relatively to the size of the Fish, and more or less degenerate, 
sometimes partially solid, but almost invariably includes two laterally situated air sacs 
with simple cavities, which together may be regarded as equivalent to the anterior 
chamber in a normal Siluroid. Lateral compartments, as a rule, are absent altogether, 
or, if present, are very rudimentary. Whatever its condition, the air-bladder is almost 
always partially or completely enclosed within transversely disposed bony recesses, 
formed either by the transverse processes of the fourth vertebra alone, or in conjunc- 
tion with those of the fifth vertebra. 
Although a convenient method of classifying morphological facts, it is obvious that 
this classification, based as it is upon so variable an organ as the air-bladder, can have 
no genetic value. 
1. Siluridm Normales. 
Under this head may be included the genera : Plotosus, Copidoglaiiis, Cnidoglanis, 
Siliirus, Wcdlago, Cryptoptcrus (certain species), CallicJirous, Schilbe, Eutropius, 
Pangasius, Bagrus, Macrones, Pseudohagi'us, Liocassis, Bagroides, Pita, Amiiirus, 
Platystoma, Piramutana, Pimelodus (certain species), Auchenoglanis, Arius, Ilemi- 
jnmelodus, Ketengus, AEiurichthys, Osteogeniosus, Batrachocephalus, Auchenipterus, 
Malapterurus, Oxyd.oras, Lais, and Aspredo. 
In the Silurolds included in this grouj) the number of rigidly interconnected 
vertebi’se varies. The-first, the complex, and the fifth vertebrae are invariably so con- 
nected together that no motion is possible between them ; and occasionally the sixth, 
the seventh, and even the eighth, may be included in the series. The firmness of 
their union naturally varies with the age of the specimen, being greater in mature or 
old specimens than in young ones, but prolonged maceration generally has the effect 
of rendering all but the complex and fifth vertebra3 more or less easily separable. 
The rigid connection of these vertebrae is effected by various means, of which the 
investment of the lateral surfaces of their centra by a continuous deposit of super- 
ficial bone, and the firm sutural union, or even partial anchylosis, of their neural arches 
and spines, are the most noteworthy. The rigidity of the complex and fifth vertebra.*, 
with the occasional addition of the sixth, may be further increased by the sutura* 
union or partial anchylosis of their respective transverse processes. The anterior 
vertebrae are also firmly connected to the skull, generally by the articulation of the 
arch and spine of the third vertebra with the exocci[)ltals and supraoccipital, the 
transverse processes of the fourth vertebra with the post-temporals, and the spinous 
processes of the third and fourth vertebrio with the supraoccipital s])ine ; less 
frequently by the formation of interlocking accessary articular processes on the 
2 G 2 
