64 MR. W. S. EOWNTKEE ON THE 



the literature of the subject to suggest that the opening is not median. It has, however, 

 long been known that in the Erythrinoids, or at any rate in the genera Erythrinus, 

 Macrodon (PI. 3. fig. 3), and Lebiasina (fig. 5), the duct opens not dorsally but laterally — 

 in fact, in the middle of the left side. Valenciennes does not seem to have been aware of 

 this peculiarity, but it is referred to by Stannius *, and has been discussed at some length 

 by Sagemehl f . No observations recording anything unusual in the other Characinidse 

 seem to have been made. In fact it is stated that these forms do not present any 

 features of interest in this connection. This I shall show to be incorrect ; and it is 

 somewhat remarkable that Sagemehl at least — uho bases an argument of some importance 

 upon the conditions found in Erythrinoids — should have overlooked the true relations 

 existing in the other Characinids. 



I have paid special attention to these relations in every form which has come under 

 my notice, for a priori it appeared unlikely that the Erythrinoids should stand absolutely 

 apart from their allies in this matter. The results are as follows : — In the Erythrinoids 

 Macrodon, Erythrinus, and Lebiasina the ductus pneumaticus opens into the middle of 

 the left side of the tract. This is seen in fig. 3, and equally distinctly in fig. 5. In the 

 fourth member of the Erythrinoid group, JPyrrhuUna, the opening is dorsal, but vrell to 

 the left side. This condition I shall refer to as " laevo-dorsal." Amongst the other 

 Characinids, out of 35 species examined by me with reference to this point, 25 were 

 found to have the opening of the duct very decidedly to the left side ; in three others 

 the opening was only slightly on the left side of the mid-dorsal line ; in the remaining 

 seven, owing to the condition of the specimens, I was not able to satisfy myself 

 absolutely, but in at least 5 of the number the opening appeared to be slightly to the 

 left. In no instance was the orifice on the right side, and in only two did it appear to 

 be median. It should be stated that in some instances only one specimen was available ; in 

 others, several. 



In Sarcodaces the duct was found to open quite in the middle of the left side 

 (fig. 2), much as in the three first-mentioned Erythrinoids. This observation was made 

 on several specimens. Thus this character, if it possess any significance, serves to 

 strengthen the suggestions yielded hj the skull of afiinity between these forms. In none of 

 the other Characinids examined does the opening lie so far to the left. Ichthyoborns (fig. 9), 

 of the internal anatomy of which no description exists, stands next to Sarcodaces in this 

 relation ; but the opening of the duct is more dorsal. Neoboriis and Eugnathiclithys are 

 closely similar. It is also w^ell to the left in Sydrocyon {brevis and/orskalii) ; in Alestes 

 [nurse, longipinnis, and macrolepidotns) (fig. 10) ; in Ilioxtlestes stormsi ; in Tetragono- 

 pterus {fasciatus and argentatus) ; in Petersius, Bryconaethiops, Salminus (fig. 11), 

 Anacyrtus, Anostomus, and Leporinus ; in ProcJdlodus (fig. 12), Xenocharax, Disticliodus 

 t^^antonii and niloticus) (fig. 13), and Citharimis macrolepis (fig. 13 a) ; and in Serrasalmo 

 (Immeralis and piraya). 



* Stannius, ' Anatomie der Wirbelthiere,' 1854, p. 224. 

 t Sagemehl, " Das Ci-anium der Characiniden," p. 108. 



