106 Observations on the Natural History 
off on each side by the trunk that springs from the heart, as 
there are branchial arches, viz. four. In the siren and axolotl, 
(which have also eight branchial arches,) M. Cuvier speaks 
only of six arteries, three on each side, going to the gills ; but 
as, by the aid of injections, we have found, say the authors, that, 
in the larvae above named, there are eight vessels, and that the 
artery which runs along the interior arch of each side, and 
which M- Cuvier has not seen, is that which in process of time 
becomes the pulmonary artery, so, guided by analogy, we hold 
it for certain, that, as the siren is furnished with eight bran- 
chial arches entirely similar to those of the other larvae, there 
are also eight arteries, four on each side, corresponding to 
them. And, proceeding on this opinion, we may remark a 
striking difference in the circulating system of the siren and 
proteus, since the artery, properly called Pulmonary, which is 
found in the siren and larvae above mentioned, does not exist in 
the proteus. Doubtless in the proteus, the air-bladder, like 
every other part of the body, is duly supplied with blood ; but 
the blood sent to it is furnished by an artery coming off, on each 
side, from one of the aortic trunks, and which artery, descending 
along the canal of the bladder, gives to it a branch, and is then 
continued to the ovary or testicle in each sex respectively. 
Besides these differences in the arterial, there are others in 
the venous system ; for the vessel which returns the blood from 
the air-bladder of the proteus, does not empty itself directly 
into the cava or the auricle, as is observed in other reptiles ; 
but into the vein which carries back the blood from the organs 
of generation, which itself enters the cava above the middle of 
the kidney ; hence in the proteus, not only the true pulmonary 
artery, but the vein also, is wanting. This anatomical fact, ascer- 
tained by repeated injections, might alone be sufficient to de- 
monstrate, that the two air-bladders with which the proteus is 
furnished are not true lungs : but as some, perhaps, may not 
yield to the force of these arguments, we shall continue the 
comparison, especially as applied to the organs of respiration. 
In the larva? of the frog and salamander, the trachea opens 
directly into the lungs. These organs have the form of two 
sacs, and, from being longer than the trunk, cannot be extend- 
ed in a straight line through it, but at the lower end are folded 
