1889.] Contributions to the Anatomy of Fishes. 327 



curved (e.g., Glyptosternum). The articular process is usually long 

 and tapering. 



Lateral cutaneous areas are almost invariably well marked and, as 

 a rule, close the distal openings of the osseous recesses in which the 

 air-bladder is lodged, but their relations to the outer walls of the 

 latter vary considerably. In some forms, the structures are in close 

 contact (e.g., Gallomystax), but more frequently they are separated 

 by a considerable interval which is usually occupied by the lateral 

 lobes of the liver, or by the anterior end of the mesonephros, or even 

 by both. 



A ductus pneumaticus may be present or absent. It is usually 

 present where the laterally situated air-sacs are connected by an 

 intermediate tubular portion, but is absent whenever the two air- 

 sacs are completely separated. 



The genus Cetopsis, referred to above, is one of the genera of 

 Siluridae in which J. Miiller denied the existence of an air-bladder. 

 In a specimen of C. candira, however, we found a rudimentary air- 

 bladder in the form of two small oval sacs. Each sac was not more 

 than 6 mm. in length, and was enclosed within the slightly dilated 

 proximal extremity of a tubular or flask-shaped recess, enclosed by 

 the transverse process of the fourth vertebra. The anterior wall of 

 the recess was perforated by a small foramen, near the complex 

 centrum, through which the tripus passed from its attachment to the 

 air-bladder to its connexion anteriorly with the scaphium. As 

 Cetopsis is the last of the nine genera of Siluroids in which J. Miiller 

 affirmed the absence of an air-bladder about which any doubt 

 remained, it is now not unreasonable to assume that an air-bladder 

 and Weberian ossicles are universally present in the group. 



Of the more general conclusions which the foregoing data seem to 

 us to warrant, we shall not now do more than draw attention to the 

 following : — 



(I.) The air-bladder and the Weberian mechanism in the Siluridas 

 are reducible to a common fundamental type, which is perhaps not 

 very dissimilar to that illustrated by the condition of these structures 

 in such normal forms as Macrones or Arius. From such a type the 

 extraordinary variations met with in different genera, or even in 

 different species of the same genus, are readily derivable, such varia- 

 tions being in part due to increasing specialisation — the result of 

 physiological causes — and partly to the effects of disuse and consequent 

 degeneration. 



(II.) The air-bladder exhibits a far higher degree of specialisation 

 in its relation to the Weberian apparatus than in any other 

 Ostariophyseaa, but this fact renders it specially liable to degenera- 

 tion, when the necessity for the exercise of its special function has, 



