48 
forming a genus at the present time are so varied that it would take 
an altogether unknown quantity of sheets of foolscap to state them. 
In point of fact a genus is a collection of species which have been 
descended from a common parent. But that collection of species may 
have undergone one or more of an infinite number of different conditions, 
and called into play those habits and structures which are most bene¬ 
ficial to that collection of individuals. It follows, therefore, that one 
set of individuals will develop one structure better than another set, 
which will not require such modification. This is exemplified on all 
sides. We find larvae of one group easily separated from larvae of another 
group. Some particular structure of the imago is different in one group 
from another, although otherwise the divergence is hardly perceptible. The 
difficulty in dealing with all these various phases is to decide how many of 
these characters are going to be considered as being ground for forming a 
genus. If say, that a species of Noctuid is closely connected with the 
genus Xanthia by its larva and egg, but in the perfect state has lost 
the general characteristics and has slightly different structures also, 
are we to form a new genus for that species ? My answer would be no 
for a genus, but yes for some sectional division. With the fresh 
discovery of minor structural divergences the number of genera are 
increasing by leaps and bounds, and it would seem that in perhaps a 
hundred years or more each species will be in a separate genus. But 
to come back to our question as to whether we should create a new 
genus for our particular species. We must satisfy ourselves whether 
the structural differences we see are oecological or not. If it is possible 
that they are so, and that the whole life-history of the insect is extremely 
close to another, and the characters of the egg and pupa (essentially 
the embryological characters) show strong affinity, there can be no 
reasonable cause for separating the two. Characters, which we have 
strong reason for supposing to be so easily acted upon, are in 
themselves of no better value than for species demarcation. Although 
in the present state of our knowledge it is difficult to understand why 
the genitalia should be so easily influenced, we are forced to assume that 
they are so, as we find that in some species that are exceedingly close 
to one another, this structure varies so much that we are able to make 
use of it as a specific character. It remains to be seen whether even 
this structure is always sufficiently constant in local races of a species. 
Has for instance Lycaena agestis from the South of France got identical 
genitalia as (what we know to be only a variety of it) artaxerxes from 
Scotland ? It is admittedly probable that it has, but if there is a diver¬ 
gence we should be compelled to give it less importance than even 
species rank. Genera at the present day appear to be made much too 
much for the sake of catalogue, and not sufficiently thought out as to 
whether they are sufficiently natural genera or not. It is also thought that 
because structural differences are used that there can be no doubt that 
the splitting is justified. Meyrick in his paper last year to the Zoologist, 
writes “ Characters of colour and general form are bad only because 
they are particularly liable to be modified by changes of environment. 
Now some structural characters are quite as liable, and are therefore 
equally bad” and then goes on to illustrate the beaks of birds. I think 
Meyrick has, however, missed out a word, and that word is generically. 
Those characters that are particularly liable, should be taken for specific 
value when possible, but if they are common to two or more species, 
