AIR-BLADDER 



225 



but in others, such as Salmo, the Siluridae, Cyprinodontidae, Per- 

 copsidae, and Galaxiidae, it opens more or less on the right. On 

 the contrary, in the Mormyridae, Notopteridae, Gymnotidae, 

 Cyprinidae, and Characinidae, it opens somewhat on the left (Fig. 

 199), and in some genera of the last family, such as Erythrinvn 

 and Macrodon, the ductus passes down to open quite on the left 

 side of the oesophagus (Rowntree [375]). 



The evidence of ontogeny seems to show that the difference in 

 the position of the opening is due rather to secondary shifting than 



ao. 



Fin. 198. 



Diagram showing the relations of the oesophagus, oe-, the pneumatic duct, p.rf, the bilobed air- 

 bladder and its blood-supply in J'rotopterus, seen from behind. ao, junction of aortic arches t< 

 dorsal aorta ; LI, left lobe of air-bladder ; l.p.a, left pulmonary artery ; p. i, pulmonary vein ; 

 r.l, right lobe of air-bladder ; r.p.c, right pulmonary artery. 



to the retention of a more primitive position in the aberrant fish 

 (Moser [304], Piper [330]). 



With regard to the original condition of the air-bladder, 

 attempts have been made to derive these various organs found in 

 the Osteichthyes from some single ancestral form. 



Boas suggested that the bladder was originally dorsal and 

 median ; that, together with its duct, it split into right and left 

 halves ; that these separated, passed round the oesophagus, and 

 reunited ventrally to open finally by a median ventral glottis. 

 This explanation, however, cannot be applied to the Dipnoi (p. 223), 



15 



