66 MR. W. S. ROWNTEEE ON THE 



supply. It follows then, if the air-bladder originated as a ventral structure, that in 

 order to arrive at the Teleostean condition the bladder and its duct must have travelled 

 round the alimentary canal, to one side or the other, until finally both attained a dorsal 

 situation. On this hypothesis, the position of the duct in Erythrinoids — and, as I have 

 shown, in Characinids generally — becomes endowed with a peculiar interest as repre- 

 senting an incomplete stage in the evolution of the Teleostean air-bladder ; as indicating 

 the path along which the bladder and duct have travelled ; and as bearing witness to the 

 primitive character of the family of fishes under consideration. 



It would appear, then, according to Sagemehl, and so far as the discussion has yet 

 proceeded, that the dorsal jiosition of the air-bladder in Dipneusti and Teleostei must have 

 been arrived at in the two cases along two different routes, lying to the right and to the left 

 respectively, each of which may have had as its starting-point the condition which is still 

 persistent in Foljiptenis. It would seem that the two cases are further distinguished, 

 firstly by the fact that in the Dipneusti the actual orifice of the pneumatic duct has not 

 travelled more than a short distance, and secondly l)y the fact that in Teleosts the 

 bladder has changed its blood-supply from the branchial artery to the dorsal aorta. 



It may be further pointed out that this interpretation of the more or less laterally 

 situated opening of the ductus pneumaticus in the Characinidse wou.ld involve the 

 derivation of the entire family from forms more primitive than Amia or any of the 

 existing Ganoids, in all the families of which I find the position of the duct to be mid- 

 dorsal. A point referred to by Sagemehl in support of his argument is the fact that in 

 Polypteriis the left sac of the air-bladder is decidedly the larger. Thus it only becomes 

 necessary to assume the complete atrophy of the already dwindling right sac. 



Against this argument, we have at the outset the facts of ontogeny, which seem, so far 

 as at present known, to represent the air-bladder as a dorsal outgrowth from the 

 oesophagus *. But this, it must be admitted, may conceivably arise from the elision of 

 antecedent stases. 



*»^ 



The discussion of this question involves the consideration of the relations of the ductus 

 pneumaticus in the other families of Ostariophyses, and even in the Physostomi 

 generally. I have accordingly extended the investigation in that direction, with the 

 following results. 



Of the Cyprinidse I have examined specimens from the five diversified genera : — 



Leuciscus [L. rutilus & L. leuciscus) (PI. 4. fig. 14), 

 Carassius ( C. auratus), 

 Barbus {B. Bynni) , 



Varicorhinus [V. beso), . 



Catostomus [C. macrolepidotus) (fig. 15). 



In all these forms the opening of the duct is laevo-dorsal, and, in some cases at least, 

 even far to the left of the median line. In fact, excepting that its position is usually 



* iliklucho-Maclay, " Uber ein Schwimmblasenrudiment bei Selachieni," Jeuaische Zeitschr. f. Naturwiss., 

 Bd. iii. 186S. 



