188 THE CLASSIFICATION OF THE OPHIDIA. 



tubular organ. Thus the inside of the tube becomes the outside, and the entire organ 

 projects freely from its base anteriorly. It finds its way into the corresponding 

 oviduct of the female, and when once in place it cannot be retracted in most species 

 without invagination. This is performed by the contraction of the now internal re- 

 tractor muscle. This is inserted on the internal face of the apex, and draws it in- 

 wards, so that it soon assumes the original ensheathed position beneath the tail. It 

 cannot be withdrawn from the oviduct without invagination, because it is generally 

 set with strong bony spines which diverge backwards. They have a perfect grip on 

 the walls of the oviduct and would, in some instances, lacerate that organ if the two 

 bodies should be forcibly drawn apart. In other cases the hemipenis would be torn 

 off at the base. Snakes sometimes partially project this organ, apparently in some 

 instances for defense, as the spines are very pungent and are sometimes curved like 

 cats' claws. Snakes are, however, very careful not to present these organs fully 

 evaginated so as to expose the delicate structures near the apex. I have never seen 

 this to be the case in an alcoholic specimen (with one possible exception), and I should 

 judge that this was the general experience, from the figures given by authors. 



The hemipenis of the Ophidia is traversed by a groove which divides the super- 

 ficial investment to the internal integument (or external integument when the organ is 

 retracted), which commences at the base internally and soon turns to the external side 

 of the organ and continues to its extremity. This is the sulcus spermaticus (s s in 

 PI. XXVII). This sulcus is always bifurcated in venomous snakes, and I find it to 

 be equally bifurcated in many harmless snakes (Figs. 2, 3, 7). The investing tissues 

 may or may not correspond with this bifurcation. Thus the hemipenis may be more 

 or less bifurcate (Figs. 1, 2, 7, 9, 10, 11). Schlegel states that it is bifurcate in ven- 

 omous snakes, but it is not so in the sea snakes Hydropliis and Uydrus, nor in Buvgarus 

 semifasciatus, Hojrtocephalus coronatus, etc., while it is bifurcate in many non-venom- 

 ous forms. Xext to the bifurcation of the sulcus in importance is the nature of the 

 surface of the external investment (internal when retracted). In the most perfect 

 types, both venomous and non-venomous, this surface is reticulate like tripe, the 

 enclosed areas forming calyces, which may have a suctorial function (Figs. 6, 9, 10, 

 11). Their borders are often papillose, and are sometimes so deeply divided into 

 papilla? as to lose their original character. These papilla? may be the seat of osseous 

 deposit, becoming bristles or spines (sp), which become larger towards the middle of 

 the length and lose their mutual membranous connections. These isolated spines may 

 extend to the apex, but they rarely extend to the base. The surface may, however, be 

 laminate and not reticulate, and the lamina? may be longitudinal (Figs. 4, 7) or trans- 

 verse (Figs. 1, 2, 3, 5). In either of these cases they may not be spiniferous. The 



