PHYLOGENY OF VENOMOUS SNAKES 51 
their respective ducti excretores. Among these different Aglyphous species 
only two seem to lack the rudimentary venom gland or the yellowish portion 
of the glandula labialis superior. 
One of the most natural outcomes of these morphological investigations 
would lead to an investigation whether the parotid secretion of these solid- 
toothed snakes is really toxic when it is introduced directly into the blood 
circulation of different animals, which are susceptible to the action of the 
parotic secretion (venom) of some poisonous species. ‘This interesting prob- 
lem attracted the attention of Bertrand and Phisalix (1894), who discovered 
that the salivary glands of two European species of Tropidonotus secrete a 
fluid which acts fatally upon guinea-pigs when injected into them. 
Still later Alcock and Rogers (1902) examined the poisonous property of 
the watery extract of the parotid of freshly decapitated specimens of Zam- 
enis mucosus upon rats and mice and of the watery and saline extracts of 
the parotid and the Harderian glands of Tropidonotus piscator upon the 
last-named animal. The injection was made subcutaneously. The parotid 
extract of the Aglyphous snakes used was a viscid mucin, quite different from 
the thin, opalescent fluid derived from the Opisthoglypha. The violent con- 
vulsions that followed the injection of Zamenis extract presented a marked 
contrast to the characteristic dyspnoeic convulsions caused by the Opistho- 
glyphous snakes, although a sufficient dose killed these small mammalia. 
The extract of the Harderian gland was without toxic effect. 
The poisonous property of the parotid gland, as well as its secretion, of 
certain harmless, non-groove-toothed snakes, has been established beyond 
doubt, but there is still much to be done about this problem, especially in 
regard to the identity of the poisonous principles of these rudimentary forms 
of the venom glands. A similar investigation of a greater number of non- 
poisonous snakes and even of the mammalia is highly desirous. 
The passing of snakes from the non-poisonous into the poisonous kinds 
is a gradual process and is associated with a definite morphological and 
functional modification of the parts directly concerned. This modification 
is an acquisition of the poisonous apparatus by grades, namely, the speciali- 
zation of the supralabial gland into a venomous one, and then the canali- 
zation of the maxillary teeth so as to enable them to conduct the venom. 
These changes are also accompanied by an ascending perfection of these 
and other accessory apparati. Thus, in one group of snakes there is neither 
the venom gland nor the poison fang. Ina second there is the venom gland 
of a rudimentary stage, but no venom-conducting tooth. In a third the 
venom gland attains larger dimensions, but the fang is still primitive, being 
moderate in size, shallow in groove, and situated inconveniently in the rear. 
In a fourth group the conditions are more favorable, as the venom gland is 
better developed, the fang is longer and has a deeper groove or a canal, 
and its position is in the anterior of the maxilla. The fifth and last group 
has a well-developed gland and one or more large and strong fangs. The 
fang is tubular and situated in the anterior part of the peculiarly short space 
of the erectile maxillary bone. 
