84 
ing apart, recalling SpAimx and Castuza, that of Attacus, Saturnia, 
and //emz/euca is, on the contrary, flowing, the veins separating, so 
that the media and its system finally entirely drop out, as such, from 
the centre of the wing, and we are reminded of the Nymphalidee by 
this circumstance. The wing of Hemz/euca prophesies the coming 
extinction of the media, that of 4g@a or Aufomerts lends itself to no 
such intention. Ag//a has gone as far as it could go in this general 
direction, and resists the giving up of a central position for IV 2. 
Leaving the neuration, I think the feeble simple female antennz 
of Aglia, Automerts, and Citheronia show an affinity to exist between 
these groups. Nor am I able to find any character to contradict my 
classification until we come to Dr. Dyar’s statements as to the larval ’ 
tubercles. Upon this question I must be silent, because to enter 
upon it fully would require material which I do not possess, and 
experience I have not acquired. I, however, by producing figures of 
the neuration, enable my statements to be controlled by any one, 
whereas Dr. Dyar’s lack this method of appeal. 
An apparent way out of this difficulty might be thought to exist in 
multiplying the families. We might take dgva as type of the 
Agliadze, Automeris as type of the Automeridee, C7theronia, of the 
Citheroniadz (this latter I am quite willing to grant). But none the 
less would the pattern of neuration teach us that the three families 
are phylogenetically related. So we might also take A/¢facus as type 
of the Attacidee, Saturnia as type of the Saturniadee, Hemzleuca of the 
Hemileucide. Equally here are the families connected. It is not a 
matter of technical classification, but one of phylogenetic arrangement. 
The Saturniades are, in my opinion, diphyletic, and the three minor 
groups on either side represent different grades of relationship upon 
the same lines. ‘lhe evidence upon which the two stems are united 
is inconclusive against equal negation. It is a matter not of cer- 
tainty, but at the most of probability. But in no case can we allow 
ourselves to bring Aulomeris and Hemzleuca together, and fasten 
them by a categorical clothes-pin, as Mr. Reuter does in an analogous 
case with the Papilionidz and Pieride.* The moment we take the 
pin away the groups fall apart. The family Aemzleucide of Dyar 
contains material which evidently has come from two different 
sources, and is, in my opinion, therefore, a classificatory freak. I must 
naturally object to Dr. Dyar’s pulling the veins about in and out like 
a telescope, to suit his theory of the larval tubercles. The estimate 
I have been inclined to place upon these latter as aids to a phylo- 
genetic classification is given in my original paper on ‘ The Butterflies 
of Hildesheim ” pp. 15, 16, and I have seen no ground to change it. 
No attempt to discredit my interpretation of the venation will, I 
* Mr. Reuter calls his clothes-pin “ Papiliones,’”’ vide ‘Ueber die Palpen 
der Rhopaloceren,” 555. The original bringing together of Automeris and 
Hemileuca in one group is due to Grote and Robinson, ‘' Lepidopterological 
Contributions,” “ Ann. Syc. N. H. N. T.,” viii, 376, 1866. But we had not 
studied the neuration. 
