58 ME. W. S. EOWNTKEE ON THE 



root of the tail-fin. The jjosterior chamber, usually much the larger, communicates 

 with the alimentary tract through the ductus pneumaticus, which is often of considerable 

 length and stoutness. These more obvious characters are to be seen in PL 3. fig. 1. 



The air-bladder consists of two coats : the outer stout, fibrous, and more or less opaque ; 

 the inner thin, delicate, glossy, and transparent. In the anterior chamber the two coats 

 are non-adherent and easily separable. In the posterior chamber they are not separately 

 distinguishable. The inner coat in the anterior chamber is thickened ventrally by what 

 has the appearance of a narrow sheet of muscular tissue running longitvidinally. In 

 front the air-bladder abuts against a stout transverse membrane which pas.ses backwards 

 ventrally to the bladder as far as the constriction, where it merges into the outer tunic 

 itself, and ends. This membrane is secured dorsally at two points, to the centra of two 

 successive vertebrfe, the first attachment taking the form of a stout peduncle ; also 

 laterally to the body-wall at two points about one-third of the distance from the mid- 

 dorsal to the mid-ventral line. At the second median attachment this membrane is 

 continent with the true outer coat of the bladder, which is also attached to the body of 

 the next vertebra. Thus there are, in all, three median attachments, the last two only 

 of which involve the true coat of the bladder. The first of the two coincides with the 

 insertion of the hindermost of the series of ossicles constituting the Weberian mechanism. 

 The description above given refers to Macrodon, in which my examination of these 

 relations went into greatest detail. But I have not noted essential differences in these 

 particulars in other forms. 



The air-bladder lies outside the peritoneum, and thus in reality ou.tside the body- 

 cavity. In the herbivorous forms Citliarinus, Dlstichodiis, and Trochilodus, and in these 

 only, this membranous partition, shutting off the bladder and kidneys from the other 

 viscera, has become tough and fibrous to such a degree as to be quite resistant to the 

 knife. It is also opaque, causing the illusion that the air-bladder is entirely wanting. 

 In the other herbivorovis forms examined, Xenocharax and Ciinmatus, the membranous 

 partition is much more delicate, but still an advance on the condition met with in the 

 other Characinidee. Here, again, it is noticeable that Xenocharax presents a lower stage 

 of differentiation tiian its allies Citliarinus and Distichodus. What significance this 

 particular modification can have is not clear to me. It is suggestive, however, of an 

 adaptation in some way either to the herbivorous or mud-eating habit as such, or else to 

 the ground-habit. It may be, and I am inclined to take this view, that it serves to 

 protect the air-bladder from pressure arising from the distension of the large mass of 

 coiled intestine. It is obvious that sucli variations of internal pressure on the air-bladder 

 would interfere with its bathymetric sensibility. 



Possibly to be connected with this feature is Boulenger's * discovery that in these 

 three genera, and these alone so far as is known, the dorsal and ventral lateral muscles 

 diverge opposite the anterior chamber of the bladder, so bringing the organ into close 

 contact at that point with the external skin. This condition is suggestive of the 

 "lateral cutaneous areas" found in many Siluroids, and stated by Bridge and Haddon to 



* ' Lea Poisaons du Basain du Congo,' p. 199. 



