LAWS OF NOMENCLATOKE. 
241 
than a thrush, should he called Tyrannus intrepidus. 
Names, on the other hand, derived from the general or 
partial colouring of the plumage, are appropriate at aU 
times ; and, together with such others as express the 
essential specific character, are perfectly unexception- 
able. Names, also, which express the local haunts of 
different species, although not so good, as being less in- 
telligible, are not to be rejectetl, as Alauda pratemis, 
Anthus aquaticus, &c. 
(200.) A genus or sub-genus can never be admitted 
into a circle which has been demonstrated to be com- 
plete, without its rank and station has been analysed, 
and is made known ; in other words, it must be proved, 
in the first place, to be natural. This rule is of course 
only applicable to such groups as have been completely 
analysed, and where the system followed is a natural 
one. We have said so much on this head in another 
volume, that recapitulation is here unnecessary. Wc 
have been inundated, particularly from France and Ger- 
many, with “ new genera,” as they are termed, which 
their inventors can give no other reason for making, 
than that they fancy them to be so. It is really time 
that we called for some better reason than this, more 
especially in ornithology; we must be pardoned, therefore, 
in future, for not giving currency to such divisions, un- 
less they are first shown to be natural. 
(201.) A genus should possess at least three positive 
and discriminating characters, hat a suh-geuus usually 
possesses only one. — It is totally impossible to lay 
down any positive rules by which either of these groups 
can be distinguished ; for they vary in almost every 
family. The above conclusion, however, is the result of 
much study ; and we must leave those who are disposed 
to verify its (ruth among some of the best known groups. 
1 1 is quite obvious, however, that a sub-genus can never 
be known from an aberrant or esculent species, until the 
circle to which it belongs has been analysed ; and there- 
fore it becomes extremely desirable, at least in a na- 
tural classification, not to mix up imaginary divisions 
B 
