54 
SYSTEM OF NATURE. 
asserts, and lie has been followed by most subsequent wri¬ 
ters, that impregnation is effected without contact; but I 
have reasons, which it is unnecessary for me to detail here, 
for believing this to be a mistake, at least in some species. 
It is sufficient for me to state that those reasons are the 
result of my own repeated observations.”* Thus we find 
this extraordinary mode of impregnation still more limited 
than that of metamorphotic respiration : it divides Rana 
and Salamandra, two genera which all authors agree in 
considering closely related. 
The great objects to be kept in view in seeking to dis¬ 
cover natural divisions, when these are not, as in the case 
of birds, obvious to every beholder, are uniformity and 
equal value: now although the characters of respiration, 
circulation and generation, appear of the greatest impor¬ 
tance, yet when we see how fractional is the portion of the 
group of reptiles to be divided from the rest by discrepan¬ 
cies in these particulars, and how vast and heterogeneous 
is the group still remaining, united by a merely negative 
character, in one instance because they do not agree with 
Rana and Salamandra, and in the other instance because 
they do not agree with Rana : he must be a bold theorist 
who shall assert that Rana, or Rana and Salamandra — or 
even Rana, Salamandra, Proteus and Siren — constitute a 
class equal to the united groups of Chelonia, Enaliosauri, 
Crocodilia, Sauria, Saurophidia, Ophidia and Amphisbaenia. 
In fact so great appears to me the objection to so unequal 
a division, that I consider this inequality alone a sufficient 
reason for doubting its value. 
Having returned this little group into the greater one 
* Bell’s British Reptiles, 122. 
