YISCEEAL ANATOMY OF THE CHAJtACINID.E. 71 



2. Even eliminating all douLtful cases, there remains a substantial minority of families 

 in which the opening of the duct is situated on the right side, the majority, however, 

 having the duct situated medially or to the left side. 



3. Many of the most primitive or generalized forms, including the three Ganoids, have 

 the duct in the mid-dorsal line. 



4. Families regarded as being allied do not always present similar conditions in 

 relation to the situation of the duct. For example, the Siluridaj and the other Ostario- 

 physes ; and, again, the Esocidte and Galaxiidoe. 



Whatever may be the interpretation of the facts, it seems clear that the conditions 

 found to obtain in the Salmonida? and Siluridae, at least, cut away tlie ground from 

 Sagemehl's argument, in so far as it was based upon the supjDOsed transitional conditions 

 presented by the Erythrinoids. His contention for a primitively ventral air-bladder, 

 which in the Teleostei has travelled round the left side of the oesophagus, may indeed 

 still be the correct view, but it can no longer be supported by reference to the Erythrinoid 

 conditions, unless with the reservation that the evidence supplied by the Salmonidse and 

 Siluridas points to a journey in precisely the opposite direction. 



If the facts of organization before us stood alone, we might legitimately deduce two 

 divergent lines of evolution for the two groups into which the Physostomi are thus made 

 to fall, eliminating those families in which the duct is median and conceivably to lie 

 derived from either condition. The idea is not without a certain fascination. But its 

 demands are surely such as cannot be conceded. 



It demands, for example, for the highly-specialized Weberian apparatus either an 

 independent origin in the Siluroids and the other Ostariophyses, respectively, or else 

 an antiquity at least equal to that of the postulated ancestral Polypteroid air-bladder. 

 The former supposition is incredible : the latter highly improbable. The association of 

 the Salmonidse with the Silurida? is a less serious difficulty, kinship not being necessarily 

 implied, but possibly only parallel development. 



If we set aside this hypothesis, as I think we must do, and admit that the duct may 

 in all cases have travelled in tlie same direction, we may perhaps conclude from the 

 weight of evidence that the i-otation has been on the left side. This, however, necessi- 

 tates the assumption that in the Siluridte and Salraonidae the duct has crossed the median 

 line and become definitely associated with the right side of the oesophagus — a proceeding 

 for which I am able to suggest no conceivable cause. Moreover, the possilnlity is not 

 absolutely excluded that the rotation may have been in the reverse direction, and that it 

 may be in the Characinid group of families that the crossing of the median line has 

 taken place. 



There still remains another view : that the immediate ancestors of the higher 

 Teleostomi had already evolved, the evidence does not show how, a dorsal air-bladder 

 connected medially with the a»sophagus, and that tlie pneumatic duct afterwards shifted 

 to one side or the other as the several families branched out. This view, however, 

 attempts no solution of the problem of the origin of the air-bladder : it merely suggests 

 an explanation of the varying conditions observed Avithin the limits of the Teleostomi. 

 Like the preceding hypothesis, it is unsatisfactory in that no cause is apparent whicli 



SECOND SERIES. — ZOOLOGY, VOL. IX. 11 



