THE CLASSIFICATION OF THE OPIIIDIA. 195 



American museums, I have not been able to investigate them fully. The great Colu- 

 brine division is remarkably constant in its undivided sulcus and abundant calyces. 

 In degenerate types the calyces become less numerous. The groove-toothed Dipsa- 

 dines have the same structure. Except one Australian genus (Acanthophis), all the 

 disciferous types are [Neotropical, and all have a double sulcus. The other Neotropical 

 types with double sulcus may be calyculate or spinous, and they present a great 

 variety of detail. Here again the glyphodont and aglyphodont types are quite 

 parallel to each other. The structure in the water snakes is again different and char- 

 acteristic. The organ is feebly spinous from the base to or near to the apex, possess- 

 ing no calyces, disc or transverse plicae, and the prehensile function is maintained by 

 one or a few large hook-shaped spines at the base. In 1864 I referred several genera 

 which had been placed in the Calamarina? to the water snakes on account of the con- 

 tinuation of the hypapophyses to the tail. I was much gratified on examining their 

 hemipenis to find that they (genera Tropidoclonium, Virginia and Haldea) present 

 exactly the characters of group to which the vertebrae indicated that they should be 

 referred. On the other hand, the characters of the hemipenis in Ablabes (baliodirus) 

 led me to suspect that it possesses the vertebral characters of the Natricinse, and on 

 examination this proved to be the case. In like manner I have been able to refer gen- 

 era supposed to belong to the Calamarinse to almost every natural division of the 

 Colubroidea by the study of the hemipenis. The old Calamarinse of authors is simply 

 an aggregation of burrowing or degraded forms of several natural groups. 



The Natricine (water snake) group is connected with the groove -toothed water 

 snakes (Homalopsinse), and both of these groups pass probably into the Lycodontine 

 series, in the typical forms of which the spines are arranged in flounces. It is difficult 

 as yet, and perhaps may not become easy, to distinguish some members of the Lyco- 

 dont group from certain ground snakes with totally spinous hemipenis, especially certain 

 African genera, as Elapops, Grayia and others. These questions remain for future 

 research. 



I have found the characters of the hemipenis as constant as those of any other 

 part of the organism. Occasional irregularities are to be looked for, but the only one 

 which I have met with is in the case of a specimen of Boaodon infernalis from South 

 Africa, in which the hemipenis is shortly bifurcate on one side and not so on the other. 

 There is a tendency to bifurcation in some individuals of Ophibolus getidus which is 

 not conspicuous in others. It is a tendency only. There are seen in many species of 

 all groups with calyces, ribs or welts having a longitudinal direction. On these the 

 calyces are crowded and closed, and they are sometimes rudimental or distorted. I 

 have not yet ascertained the constancy of these structures in species and genera, 



