18 
probably be that one of the nearctic “ species ” is really the closest 
relative of our trifasciata in which case ruberata, too, will become an 
outcast from Hypsipetes , of which trifasciata is the genotype. 
No one has pressed this position to its absolutely logical conclusion 
as yet, though Mr. Tutt has come rather near it in many of his recent 
revisions. Is the tendency to be encouraged and developed ? If we 
want so many units, why not take the species itself as a sole unit, 
establish — as I have already suggested — a mononymic usage through¬ 
out the order Lepidoptera, and devise your biological groupings with¬ 
out imposing the almost unbearable burden of interminable generic 
names ? If a genus is a “grouping of convenience,” can we say there 
is any greater convenience in recognizing that only trifasciata and 
ruberata are to be called Hypsipetes, than that a larger related 
assemblage shall be so called ? Manifestly, from the point of view of 
the frailties of human memory, it will be a greater /^convenience. 
But the defence would not go along this line at all; and every earnest 
worker will concede that some sacrifice of mere convenience is often 
demanded if it enhance the cause of truth. 
The arguments in favour of the recognition of small genera are, I 
take it, two, and two only; and I think I shall be able to show that 
both are fallacious. 
(1) It is argued that to provide the “next grade above the species” 
with a separate name, and to use that name in binomial conjunction 
w 7 ith the species name itself, gives the clearest and most up-to-date 
insight into questions of relationship; it being assumed that the higher 
grades of relationship will be indicated by tribal, subfamily and family 
names, and so on. But we cannot, I think, provide all these grades 
with names, and — as I shall show almost immediately—we are told 
we must not use even one of the grades which might help us, the 
subgenus. Barring the subgenus, the most which we can summon to 
our immediate serviee, so far as I am aware, will be: subtribe, tribe, 
? supertribe, subfamily, family, superfamily. I have been trying to 
think how far these would go with a large, half-natural and yet very 
diversified genus like Boaniria, Tr., Hmpsn. We cannot raise it 
higher than the subfamily rank — Boanniinae; for it is certainly part of 
a larger group (including Nychindes, Synopsia, etc., etc.), which would 
have to be a family— Boarmiulae, —leaving the Geometers as a whole 
to become the superfamily. If we allow for “ supertribes ” (rather an 
innovation) we may possibly just get through our scheme, lor this 
supertribe might limit us to the fairly typical Boannias, such as are 
brought together generically by Staudinger and others, leaving out¬ 
side some of the more heterogeneous oriental elements which Hampson 
has included as subgenera. The two “tribes” would then, perhaps be 
Boarmidi ( = genus Boarmia of our British lists) and Ectropidi ( = genus 
Ectropis, Hub., = Tephrosia of our British lists) and we should still have 
a little room to disport ourselves among the inner relationships, as for 
instance by separating biundularia and crepuscularia generically from 
the rest. But what are we to do if we should discover that consonaria 
is nearer to these two than this group of three to luridata ( e.vtersaria) 
or to punctularia / Shall we try' - the “supergenus?” I need not 
labour the point further. The real answer to the argument does not 
lie in the almost impracticable nature of the scheme, but in the fact 
xix. 
