Cope.] -^1" [May 18, 



characters derived from the skeleton define only the larger divisions : 

 ■while those derived from the teeth are not sufficiently numerous and 

 important, vpith some exceptions, to indicate all of the affinities of the 

 genera. I think I have added materially to the means at our disposal for 

 classifying the Ophidia, by a study of the hemipenis.* The criteria will 

 however not be sufficient until the entire anatomy is worked out, and in 

 the present paper I add what may be found by a study of the lungs.. It 

 seemed probable to me that an organ which presents so much variation, 

 as above stated by Stannius, must furnish some important clues My 

 examinations embraced about one hundred and fifty species, which 

 represent nearly all the leading types. The results are presented in the 

 following pages. It will be seen that they correct the statements of 

 Schlegel and Stannius in a number of important respects. 



The snakes with rudimental posterior limbs (Peropoda), show in the 

 character of their lungs, what they show in the rudimental limbs them- 

 selves, and in the hemipenis, the nearest relationships to the Lacertilia. 

 They possess, with an exception to be noted later, two well-developed 

 lungs, one of which is larger than the other. The smaller lung lies to 

 the right side and ventrally, while the larger one lies to the left side and 

 dorsally. In some species the dorsal and ventral relation is more pro- 

 nounced than in others. In the Colubroidea the right or ventral lung is' 

 generally present, but of very much reduced proportions, the usual size 

 being from two to five millimeters in length. It is connected with the 

 other lung by a foramen which perforates the tracheal cartilage at a point 

 a little beyond the apex of the heart, and opposite to the proximal part of 

 the dorsal lung. It is sometimes connected to the dorsal lung by a short 

 tube, in which cartilaginous half rings are seen in but two of the genera 

 examined, viz., Heterodon and Conophis. The lumen of the rudimental 

 lung may be lined by the same reticulate structure as is seen in the dor- 

 sal lung, or its waHs may be smooth. In sonie Colubroidea the rudi- 

 mental lung is absent, but such species are relatively few. 



The dorsal lung may present proximally alongside of the trachea an 

 auricle or pocket, and this is so developed in the genus Heterodon as to 

 reach to the head, without communication with the trachea, other than 

 that furnished by the normal portion of the lung. In the Solenoglypha, 

 without exception, this extension of the dorsal lung is present, and extends 

 to the head, and its lumen is continuous with the trachea throughout its 

 length. The same structure exists in the genera Il3'drus andHydrophis ; 

 and also in the peropodous genus Ungualia, whicli differs besides from other 

 Peropoda in having but one posttracheal lung. Finally the tracheal lung, 

 as I shall call it, is distinct from the true lung inPlaturusand in Chersydrus. 

 In the former of these genera the trachea is not separate from the lumen ; 

 while in Chersydrus it is distinct. It, however, communicates with the 

 cells of which the lung consists in this genus by a series of regularly 

 placed foramina on each side. There is no lumen in the tracheal lung in 



* American Naluralis!, 1893, p. 477. 



J 



