228 THE AMERICAN NATURALIST. [Vou XXXIII. 
mals, we find an articulation of stapes with quadrate, while in 
the reptiles no such articulation occurs. 
Another fact which has a bearing upon the question of the 
origin of the mammalia has not, so far as we are aware, been 
referred to. In the urodeles! there develop a pair of thoracic 
ducts, one duct emptying into the venous system on the right 
side, the other into the left, near the heart. Of these ducts, 
the left is from the first the larger, while a little later the right 
completely disappears, leaving the left as the functional duct of 
the adult. This disparity in size from the first would show that 
this left-sided condition had persisted for a long time. In the 
mammals, as is well known, the left thoracic duct alone is 
functional, while in all sauropsida it is the right that persists. 
It is, therefore, impossible to derive the mammalian conditions 
from those found in any existing reptile, but of course one can- 
not say but what the earlier reptiles had this part of the lym- 
phatic system paired. However, the conditions in the urodeles 
are suggestive. 
We would not be understood to derive the mammals from 
any true urodele stock, but from some ancestor not widely 
removed from them. The urodeles, as we know them to-day, 
are a degenerate group, possibly descendants from terrestrial 
forms, and the lowest of the group, like Necturus, etc., have 
departed most widely from the ancestral type. The urodeles 
have lost many cranial bones; they have reduced the ribs, they 
have lost entodermal gills and gained those of ectodermal origin ; 
they have lost the Eustachian tube, but they have retained 
many features which make them extremely interesting in con- 
nection with all phylogenetic speculations. 
If, now, we advocate the amphibian origin of the mammals, 
we must consider the arguments of those who would derive 
them from the theromorphous reptiles. The chief of these 
are as follows: 
In certain theromorphs, as in most mammals, there is a het- 
erodont dentition ; incisors, canines, and molars being differen- 
tiated. This, however, is not conclusive, since a heterodont 
1 For our knowledge of the 1ent of thel tem of the urodeles 
we are largely indebted to the sor inveitigntions of Dr. F. D. Lambert. 
