94 
Annals of the Transvaal Museum. 
and forming the anterior wall of the orbital cavity, but in Bitis arietans 
and Crotalus horridus the orbital portion is bent forwards and upwards 
away from the orbit forming a large and broad bony mass more or less 
parallel with the long axis of the skull. It should be mentioned, however, 
that even in the same genus the prefrontal character may show a fairly 
wide range of variation, and in Bitis utropos , for instance, the orbital 
portion of this bone, though much thickened, is directed, not horizontally 
forwards, but obliquely downwards. In Ccmsus rliombeatus the special 
viperine features are ill-pronounced, for the prefrontal is precisely of the 
colubrid type and the maxilla is a three-sided bar obliquely arranged, its 
posterior side merely in apposition with the antero-inferior portion of the 
prefrontal without forming any such well-defined joint as occurs in Bitis. 
It is significant that in other respects the genus Causus is of a more 
generalized type than ordinary vipers, for the scutellation of the head is 
quite like that of the Colubridae, whereas in most vipers the head shields 
are more or less broken up into small scales. The genus Atractaspis, 
which as regards size of fangs and reduction of the palatal and mandibular 
teeth is very specialized, has a type of maxilla representing an early stage 
in the evolution of the Bitis type, for it is simply a short block of bone 
lying immediately below and movably jointed with the prefrontal, but as 
a whole showing no vertical elongation. The head scaling of this genus 
also is of the colubrid type. The evidence of this snake should, however, 
be received with caution, as probably it is a downgrade type, for the skull 
has the same degraded features as occurs in other genera of burrowing 
snakes (cp. Prosymna, also Silybura and Typhlops) ; moreover, such 
snakes (cp. Prosymna, also Macrelaps, which have fewer teeth than any 
other S. African opisthoglyph) are apt to show a greatly reduced dentition, 
accompanied by an unusual enlargement of one or two individual teeth, so 
that the very large fangs of Atractaspis may be explained in the same way, 
and, further, the well-developed maxillo-prefrontal joint considered in 
conjunction with the slender transverse bone which is hardly strong 
enough for the viperine function — observations on the jaw movements in 
life are a great desideratum — suggests that this snake is derived, not directly 
from the Colubridae proper, but from a more normal viper. 
Amongst the Proteroglypha there are many genera (cp. Dendraspis) 
in which the poison fangs are indistinguishable from those of vipers ; in 
many Cases (Naia, Sepedon) the maxillary bone terminates anteriorly 
immediately below the prefrontal, whilst posteriorly it is much shortened 
and may be entirely without solid teeth. Further, in addition to the 
shortening of the maxilla, there is generally an enlargement, especially in 
a vertical direction, of the fang-bearing portion, so that the bone is 
carried upwards along the anterior surface of the prefrontal. Such a case 
is that of the genus Naia, which in the general relation of the fang-bearing 
portion of the maxilla to the prefrontal bone shows a resemblance to 
Causus, but differs in that there still remains posterior to the fangs a short 
but almost edentulous portion of the original maxilla. Sepedon haemachates 
presents a further stage in that the maxilla has entirely dispensed with its 
solid teeth, though a short length of toothless bone still remains. 
A complete series of intermediates between the type represented by 
Naia or Sepedon and that of Causus does not exist apparently, but this is 
no serious objection to the view here presented, for the gap is not a great 
one, and, moreover, a possible explanation is suggested in the special 
movement of the maxillary apparatus in vipers : as soon as the evolutionary 
process had reached the stage when it became possible and advantageous 
