72 ME. W. S. EOAVXTEEE 0\ THE 



could have determined a definitely-directed wandering of the duct away from tlie median 

 Hue ; and it is to he remenihered that in some forms — uotahly the Erythrinoids — the 

 deviation is very great. 



Of the three hypotheses here advanced, the third is most in harmony with the con- 

 dition met with in Amia and the other Ganoids, and, indeed, as it seems to me, with the 

 facts generally. As regards, however, the testimony afforded by the Ganoids and other 

 so-called " primitive " forms, it is sometimes necessary to remind oneself that tliey are 

 themselves final terms, not mean terms, of series which have survived to the present 

 dav, and that, whilst retaining certain primitive characters, they are not on that account 

 necessarily less likely to have acquired extreme modifications in other directions. 



With reference to the jiossible causes which may have operated in determining the 

 position of the duct, it may be pointed out that in some forms — notably perhaps in 

 Amia and Lepidosteus — the ductus pneumaticus is so short that any asymmetry in its 

 position would almost inevitably he attended by an asymmetry in the position of the 

 air-bladder itself. One can thus see a determining cause for its median iiosition. 

 In many Physostomi, on the other hand, such as the Characinidse, for examjile, the 

 duct is of such length and laxity that its asymmetry would appear at the present day to 

 be independent of the position of the bladder, whatever may at one time have been 

 the case. Whilst, however, it is fiiirly obvious that in many cases the precise position 

 of the duct in relation to the median line may be physiologically immaterial, we do not 

 thus get any light upon the cause, either of its u.uiform asymmetry throughout entire 

 families of fishes, or of the two opposite phases of asymmetry which characterize 

 different families. 



The one clear outcome of the investigation which I wish to emphasize, is the 

 demonstration that Sagemehl's interpretation of the Erythrinoid condition of the ductus 

 pneumaticus is not justified by a more extended knowledge of the facts. That in- 

 terpretation may, indeed, vet prove to be the true one — the asymmetric condition of the 

 air-bladder in Polijpterns sufiices to lend colour to it ; but it cannot be held to be 

 established by Sagemehl's line of argu^ment. For it must be remembered that the 

 argument was to a large extent based on the improbability of a wandering of the duct 

 from the mid-dorsal line having occurred ; and it is abundantly evident from the facts 

 set forth in this paper that exactly such a wandering must have taken place, to one side 

 or to the other, or perhaps to both. 



The vieAV that the air-bladder originated as a ventral structure remains, however, 

 untouched. The blood-supply of the bladder in the Dipneusti would appear to be 

 conclusive on that point. But as to when and how the air-bladder became a dorsal 

 structure in the evolution of the Teleostonii, evidence is as yet wanting. 



In comparison with the considerations just discussed, observations on the relative 

 length of the ductus pneumaticus carry but little interest. I may, however, state, 

 as the result of careful measurements, that its length in Characinids is commonly about 

 one-tenth or one-twelfth of the length of the body of the fish. Considerable divergences 

 from these proportions occur in certain forms. Thus in Macrodon the length found was 



