686 ï>' HAHKIS H WIIJ)ER, 



where anteriorly the lateral partition has effected a complete sepa- 

 ration between the two cavities. 



Siren exhibits a slight variation from the stage represented by 

 the Axolotl. Here the accessory cavity lies medially instead of laterally 

 with respect to the main chamber. Anteriorly, also, the accessory 

 cavity is again divided and ends in a double blind-sack. The medial 

 position of the accessory cavity would represent, according to the 

 Sarasins, a more primitive condition, inasmuch as they have shown 

 in the embryology oï Epicrium that the accessory nose lies originally 

 medial, and by a secondary torsion arrives at its lateral position. 



The matter, however, may hardly be considered solved, and until 

 we know more of the embryology of the Urodelans on this point, we 

 cannot express a decided opinion. The fact that in this form the 

 anterior end of the accessory cavity is forked, I consider of minor 

 significance. 



The progressive steps in complexity in the Amphibian nasal cavity 

 may be represented by the following diagrams, each representing 

 a cross-section through the left nasal cavity of the respective animal, 

 the medial side being towards the right. 



Proteus 2'riton Plethodon 



Siren Epicrium 



This subject has been treated and a similar series presented in 

 the fourth volume of the work by the Drs. Sarasin in the Ceylonese 

 (iymnophiona. Although the phylogenetic series given by them corre- 

 sponds almost completely with mine, I will yet present the above, 

 partly because I had prepared it previously to the appearance of the 

 volume just cited, and partly because of the addition of the most un- 

 important links which Siren and Siredon present. The conclusions 

 which the Drs. Sarasin reach, that the accessory cavity is the homo- 



