THE CLASSIFICATION OF Tlllfi OPHIDIA. 193 



Academy I published, in January, 1895, descriptions of the penial characters of sev- 

 eral genera which I had not previously observed. The present memoir presents a 

 number of modifications of the system as proposed by me in 1893. I have ceased to 

 regard the more important penial structures observed as definitive of families, but 

 rather of subfamilies. I hive come to regard the flounced structure as of less impor- 

 tance than at first appeared, and I find it to be characteristic of genera and groups of 

 genera only. On the other hand, I find the disciferous type to be quite distinct from 

 all others and distinguish by it two subfamilies, the Xenodontinse from the Dromicinse, 

 and the Erythrolamprina?. from the Philodryadinae. I have combined the supposed 

 Pseudaspidinre with the Lycodontinre. I have found the genus Chrysopelea to resem- 

 ble the Dipsadinae more closely than I had at first thought and have cancelled the 

 supposed subfamily Chrysopeleinae. 



In preparing this memoir I have examined material belonging to the Museum of 

 the Academy of Natural Sciences of Philadelphia, to the United States National 

 Museum, and to the Philadelphia Museums, to whose officers my thanks are especially 

 due. I wish to acknowledge also my indebtedness to Prof. Alexander Agassiz for the 

 opportunity of examining some Australian species; to Prof. Charles S. Dolley for a 

 collection from Hainan, China ; and to Prof. Wright, of Oberlin, O., for a small but 

 valuable collection from So ,th Africa ; to Mr. George K. Cherrie for a fine series 

 from Costa Rica, and Drs. Ferrari-Perez, Bernad, Dugds and Yillada for Mexican 

 species. To Messrs. J. B. Wood and George Pine I am indebted for collections from 

 Florida ; to Prof. W. T. Cummins for material from Texas ; and to Dr. Jos. Corson, 

 U. S. A., for specimens from Mobile, Ala. I am also especially indebted to the 

 Zoological Society of Philadelphia and its Superintendent, Mr. Arthur E. Brown, for 

 specimens from the Gardens. 



IV. SYSTEMATIC CONCLUSIONS. 



Diversity of lung structure accompanies the primary groups which are character- 

 ized by peculiarities of the skeleton to such a degree that we are warranted in accord- 

 ing it a high systematic value. Thus angiostomatous and peropodous snakes have 

 two lungs, while the Colubroidea have one and a rudiment, and the Solenoglypha 

 always have a tracheal lung. Exceptions and variations from these rules thus become 

 of importance. Thus I have no doubt of the propriety of the separation of the 

 Ungualiidae from the other Peropoda, on account of its pulmonary characters. Nor 

 is there any doubt in my mind of the necessity of the separation of the Leptognathinre 

 from the Dromicinse, on account of its large tracheal lung. The very marked char- 

 acters of the genus Acrochordus characterize the family, as well as the osteological 



