70 ME. AV, S. EOWNTREE ON THE 



The foregoing analysis may be thus tabulated : — 



Fishes with 

 air-bladder 

 communi- 

 cating with 

 alimentary 

 canal by 

 a ductus 

 pneumati- 

 cus. 



VENTRALLY. 



Opening- Opening a 

 median. little to rig-ht 

 Bladder yen-i side, 

 tral : its left' Bladder 

 sac the larger.' dorsal. 



Communicating Polypterus. 



with oesophagus 



or the part 



of the stomach 



immediately 



behind it. 



Communicatingi 

 with the stomach 

 in its middle 

 or posterior 

 region. 



Dipneusti*. 



DORSALLY. 



Opening on 

 left side. 



MormyridsB. 



Notopteridae. 



Characinidae. 



Gymnotidie. 



Cypriuida3. 



Opening practically median. 



Inclined 

 to left. 



Inclined 

 to right. 



Esox. 



Acipenser. 

 Aniia. 

 Lep'tdosteus. 

 Osmerus. 



Elops. 

 Albula. 

 Chtpea. 

 Chirocentrus . 

 Ualosaurus. 



Hyodon. 

 Coregonus ? 

 ThijmaUus. 



Opening on 

 right side. 



Sahno. 

 Siluridte. 

 Cypriuodontidse. 

 Percopsidee. 



Galaxiidee. 



Now, what is the meaning of the facts shown in the above table ? 



The constancy of the conditions in the families most extensively examined is strong 

 evidence that these relations are not without some significance — not a mere anatter of 

 chance. In no single genus, and indeed in no single family, have I found the duct 

 varying between right and left. AVithin any one family it may be median or more or less 

 displaced to the left side ; or, on the other hand, it may be median or more or less disjDlaced 

 to the rio-ht side. Beyond these limits 1 liave not found any variation. Throughout the 

 whole series of observations I have endeavoured to guard against fallacy arising from a 

 mere mechanical twisting of the stomach on its axis. Such a twisting, it is true, does 

 sometimes api^ear to exist either as a normal or an occasional condition ; but the torsion 

 produced does not involve the oesophagus in such a way as to lead to erroneous observa- 

 tions, except perhaps in such a case as Galaxias, where the duct communicates with the 

 stomach near its extremity, and on what appears in relation to the pylorus to be the 

 rio-ht side, but in which the pylorus is itself directed to the right. 



Looking now carefully at the table before us, we may, I think, note the following 

 fairly obvious points : — 



1. The position of the ductus pneumaticus in relation to the alimentary canal is a 

 characteristic feature in at least some groups of fishes, and as such may have a certain 

 diagnostic or taxonomic value. 



* Giinther, A. : " Description of Ceratodvs," Phil. Trans., vol. 161, 1871. 



Parker, W. N. : " On the Anatomy and Physiology of Protoptervs annectens," Trans. Irish Acad., vol. xxx. pt. 3. 

 Spencer, B. : " Contributions to our Knowledge of Ceratodus — Pt. I. The Blood-vessels," Macleay Memorial 



Vol., Linn. Soc. N.S.W. 

 Spencer, B. : *' Der Bau der Lungen von Ceratodus und Protopterus," Zoolog. Forschuugsreisen in Australia 

 und dem Malayischen Archipel, Jena, 1898. 



