60 
that the divergence of a group of species took place more recently than from 
the family ancestor, we are correct in placing those species in a genus. 
Although we have departed from the idea that the families could be 
placed as linear descendants one of the other, we have hardly thrown 
off the old idea when we come to genera in a family. But it is equally 
fallacious. Let us consider the special Leucanias again. It is 
not the existing external feeders that have been derived from the present 
internal feeders. That the two are derived from a common stock is 
true, but how many generations are we to include for the next sub¬ 
division ? One thing always seems to me clear, that if we are to sub¬ 
divide so greatly it will be necessary to have many more terms for those 
subdivisions than we already have. Our genera hold too varying 
a value. At any one period of the world’s history it is quite impossible 
to know what the relations were then of species to species and groups 
to groups. We have no means of ascertaining when one variable insect 
gave place to those varieties which of themselves were destined to 
become species. Some took place at one period, some at another. 
Those that took place earliest have therefore become more fixed, and 
those the latest least. But must we have genera so unequal that those 
that have remained as genera, so called, for ages, are to stand alongside 
small groups of species (also called genera) that have a less general 
relationship, and therefore different relationship ? Let me try to illustrate 
my point. For many years we included all our five Theclids in one 
genus Theda. They were acknowledged to be related by certain ties. 
With more intimate knowledge of our species it was found that we 
could subdivide our genus into three smaller groups, and we called them 
Zephyrus, Callophrys, Theda. But Theda has altered its relationship, 
and, instead of being a genus, has moved up to a family relationship in 
reality. But we still retain it as a genus and leave that family 
relationship or higher, wider grouping without any designation at 
all. The old arrangement was, if the species were really related, 
perfectly sound, perhaps not so convenient, but there was no error. 
We make the error in our supposed rectification in not giving Theda 
a family relationship over the genera into which it divides up. 
Attempts have been made to cope with the difficulty and we find genera 
split up into named sections, a course which gives a genus a general 
relationship to the subdivision, and is equivalent to calling if a family. 
We have the old genus Acronycta thus split up. But it is quite possible 
that some day we shall find that one of the sections is subdivisible, 
and then it would be necessary to find some new name for the value of 
that subdivision. The greater this subdivision the more relationship 
is it necessary to give if we wish to avoid artificial methods for locating 
our species. 
One more word about genera. From what has already been said 
it should be clear to everyone that we at present have genera that can 
be said to stand for father, grandfather, back to an almost indefinite 
relationship to the collection of individuals under it. In just the 
same way as species in a genus, all the genera of a family, countless 
ages back, are descended from a still more remote ancestor. If we 
could only have the written pedigree of every insect we at present 
know of, we should require an almost endless number of headings, in 
other words relationships. It thus always, to my mind, appears that 
it is quite impossible to really naturally classify with only the few 
