THE CLASSIFICATION OF THE OPHIDIA. 203 



examining but one individual, 1 am not sure how constant it is. In the Trojridocloaium 

 liaeatum, where a similar character is present, I have found it to be entirely constant. 



Boulenger distinguishes two principal divisions of ground Colubridse as genera, 

 under the names Zamenis and Coluber, on dental characters. In the former the max- 

 illary teeth increase in size posteriorly, while in the latter the posterior teeth are not 

 longer, and may be shorter than the anterior. That this distinction is valid in many 

 instances is well known, but it is admitted by Boulenger that in other instances the tran- 

 sitions are complete. An examination of the penial characters leads me to the opinion 

 that each of these groups is a series of genera rather than a single genus. Thus in 

 the Zamenis gemonensis, the type of the genus, we have the normal colubrine struc- 

 ture, from which two divergent lines may be traced. In one of these, represented by 

 the Z. ventrimaculatus, the calyces preserve their character, but the few papillae are 

 ossified as acute spines, the character defining the genus Acanthocalyx. In another 

 direction the walls of the calyces are thickened and support several series of pajiillse. 

 This is seen in the' Z. ravergierii. In the next type these numerous papillae are ossi- 

 fied, giving us the genus Gonyosoma. A greater modification is seen in the Z. jioru- 

 lentus. Here the thickening of a part of the calyx walls is greatly increased, while 

 other walls, including all of the longitudinal ones, disappear. The result is a mass of pap- 

 illose pads, a character quite different from anything else in the order, and one which 

 defines the genus Tylanthera. The explanation of this structure is rendered possible 

 by that of the Zamenis ravergierii (PI. XVI, Fig. 4). 



The North American species referred to Zamenis by Boulenger have been sepa- 

 rated under the name Bascanium by Baird and Girard. Most, if not all, of these 

 species differ from the typical Zamenis gemonensis in possessing one or two large 

 hooks at the proximal part of the spinous tract (PI. XVIII, Fig. 1) which remind one 

 of the Xatricinse, and which are not found in the typical forms of Zamenis.* The 

 Drymobius pulclierrimusf Cope possesses a similar peculiarity, which separates it from 

 the typical species of that genus. It differs from the species of Bascanium, however, 

 in having the large spines distad to the spinous tract and not proximad (PI. XVIII, 

 Fig. 4). 



In the species of Coluber there are distinct naked tracts or bands extending more 

 or less downwards from the apex (PL XYI, Fig. 2; PI. XXI, Fig. 3). There is one 

 strongly pronounced in C. cmoryi and there are two less extensive in C. dbsohtus. 



*This character is present in B. flagelliforme, B. laterale, B. mentovarium and B. mexicanum (Zamenis D. & B.). 

 In B. constrictor it is sometimes present and sometimes absent (see PI. XIV, Fig. 6). 



•(■The species figured and described by Bocourt {Miss. Sci, Mexique, p. 725, PI. XLIX, Fig. 3) under this name is 

 quite distinct. It is much larger, has but two preoculars, four bands instead of two, and is olive and brown instead 

 of black and white with a green head. I propose to call it Drymobius lemniscatus. 



