41 
the frenulum is present. We can thus make a division, just 
before those groups that have not yet lost the frenulum. We are in 
this way able to make an artificial line between butterflies and moths, 
and as a matter of convenience it is justifiable. All the species that 
fall under the head of Rhopalocera, form a very homogeneous group, 
but at the same time there is no great gap between this group and 
Heterocera. There is no reason if we make one large subdivision, why 
we should not form other large subdivisions. We might for instance 
invent a name that includes Arctiidae, Notodontidae, Lymantridae , 
Hypsidae and be equally justified. The bulk of our attention, however, 
has been in the past few years to get much smaller subdivisions in 
order, and indeed it is essential that we must know the family relation¬ 
ship before we can group families under still more comprehensive 
headings. 
It is to the nature of families and their relation one to another 
that a large part of the work has been directed. What are the 
conditions for a family ? Asked point blank in this way, probably 
a dozen individuals would each have a different answer varying with 
their idea of the value of different structures. As in the case of a species 
it is often a matter of opinion, so it is too frequently with families. 
There cannot, however, be any duplicate methods of grouping in 
nature, and for so-called utility, we should avoid any method which 
is not strictly with the laws of the universe. But I think that families 
have their being pretty generally in this wise. Species that have any 
one or more structural characters essentially the same in ovum, larva, 
pupa and imago, belong to a family. At the present time there are 
some workers who hold that the characters of the perfect insect have 
preponderating value over those of the other stages of the insect’s 
existence, and we hear of such and such insects being relegated to the Geo- 
metridae because the neuration of the wings is of Geometriform arrange¬ 
ment, or again, that an insect is a Noctuid, because of its having neura¬ 
tion that holds for that family. This shows undoubtedly that if the two 
arrangements clash, there is something radically wrong with ourwork. It 
is inconceivable and absurd that an insect can belong to different families 
at different periods of its existence. The knowledge of the life-histories 
of thousands of exotic species will prove whether the work of Hampson 
and others, who cling so persistently to neuration, is on a sound basis. 
If it be finally proved that the system of grouping by neuration 
is valid, we have one of the most useful discoveries in the whole annals 
of entomology. In the second place the utility of such a system is 
immense. As already stated, there are thousands of exotics and even 
many others besides, of which nothing is known save the perfect 
insect. Of the value of neuration as a set off to other imaginal 
characters, we rely on its immutability by external circumstances and 
surroundings. 
The antenna;, if our interpretation of their use be approximately 
correct, would not lend themselves for such a use. We suppose for 
instance that all the wingless females or semi wingless females once 
had better developed wings, and that, through disuse, they have 
gradually become atrophied, and in some instances have actually 
been lost altogether. As a consequence of this, I should be in¬ 
clined to consider that the antennae of the S s (if their use as 
scent organs is accepted) have undergone specialisation, be it in size, 
