VISCEEAL ANATOMY OF THE CHAEACINID^. 65 



In the following forms the duct opens, according to iny observations, only slightly to 

 the left of the median line : — Alestes Kotscluji, Citharinus {latus and Geoffroiji). 



Finally, Avith regard to Tetragonoiderus {abramis and multiracUatus), Curimatus, 

 Nannocharax, and Phago, I can only state, with some hesitation, that the opening 

 appeared to be slightly to the left ; and with regard to Taraphago and Micndestes 

 acutidens, that it appeared to be in the mid-dorsal line. In these latter instances, however, 

 either the size or the condition of the sj^ecimen before me was always an obstacle to an 

 accurate determination. Viewing these results as a whole, I can state with contidence 

 that, in spite of some individual variations, the position of the duct is fairly constant 

 specifically, and is normally on the left side throughout the Characinidae, so far as my 

 observations have extended. The meaning of this asymmetric structural condition has 

 been discussed at length by Sagemehl, who, however, as already stated, only recognized 

 its existence in the Erythrinoid division. According to the view set forth by that writer, 

 this condition su^iports the argument for a primitively ventral air-bladder, and is 

 inexplicable on the hypothesis advanced by Boas * of the dorsal origin of that organ. 

 The matter is of sufficient importance to warrant a summary here of the argument set 

 forth by Sagemehl, and the fact that my own observations, as will be seen, have an 

 important bearing upon the question must be my justification for venturing to 

 restate it. 



Discussing the homology of the air-bladder with the ventrally-lying lungs of higher 

 vertebrates, Boas contended that the primitively dorsal and impaired organ must have 

 undergone longitudinal fission, together with its duct ; and that the now separate ducts, 

 with their orifices, must have travelled round the oesophagus until they merged in a 

 common ventral opening, producing the condition existing in Pohjpterm and in air- 

 breathing vertebrates. 



Sagemehl objects to this hypothesis on the following grounds : — 



(1) The absence of any imaginable cause for the occurrence of the changes involved ; 



whereas a wandering in the reverse sense would naturally result from the 

 acquisition of a hydrostatic function by a ventrally situated bladder. 



(2) The absence of any satisfactory explanation of the conditions found in 



Erythrinoids and in the Dipneusti, in which the single duct passes to one side 

 of the oesophagus. 



(3) The i-elations of the lung-arteries in Ceratodus t, which arise on each side from 



the fourth branchial arch, that from the left arch passing under the oesophagus 

 and following the ductus pneumaticus, which in the Dipneusti winds round to 

 the right. If the dorsal situation of the bladder were the primitive one, the 

 arteries on each side wovdd pass directly backwards to their distribution, as 

 those on the right side actually do. 

 This last objection appears to be conclusive as regards the Dipneusti, and inferentially 

 for all dorsal air-bladders, although in Teleosts the bladder receives a different blood- 



* Boas, " tJber den Conus arteriosus uiid die Artericnbogen der Amphibicu,"' ilorph. Jahrb., Bd. vii. 1881, p. 566. 

 t Boas, " Herz uud Arterienbogen bci Ceratodus uud Protopterus," Morph. Jahrb., Bd. vi. 1880. 



10* 



