MEMOIRS OF THE NATIONAL ACADEMY OF SCIENCES. 81 



DioptidcTB (inchuliiig Phrygauidia) aud of the C.yllopodidai is nearly identical with that of 

 Melauchroia, and the latter is a true geometrid in its venation, and in the shape of its larva, being 

 a looper. Of its pn]>a we know nothing. The venation of the geonietrids is very persistent 

 Hence I couclnde that the day-tlying, nsually bright-cokned Dioptida- and Cyllopodid^, as well 

 as the Uypsida', are direct offshoots from the Lithosian stem, and that their general resemblance 

 to such Lithosians as Crocota aud Endule, as well as Ameria, is based on real afiduity. The day- 

 flying habits of souie geonietrids is also well kiujwn. The larva of Euphanessa is a geometrid, 

 but its moth has been usually associated with the Lithosiida', though its venation is geometrid. 

 Riley describes the larva of tBtoff «)•(■« as having "extremely small" anterior abdominal legs, the 

 anal ones being much longer. Probably when we learu more of the transformations of the families 

 ■we have mentioned it will be found that the in-esence or absence of certain abdominal legs will 

 l»e found to be a secondary adajitational character. It is noticeable that the dull-colored 

 Phryganidia, with only incipient clouds instead of bars and spots, is a primitive form as regards 

 markings. 



After an examination of the pupal and imaginal characters of Geometrids, Dioptids, Hypsids 

 and Syntomids, it seems to me that all these groups rei>resent more or less jiarallel lines of 

 development which originated from the generalized Lithosiida', the latter, with the Zyga-nida', 

 having sprung from generalized Tiiieina. The Nolidic represent a side branch, which evolved 

 Ironi a Lithosian perha])s like Clemeusia. The Arctiida^ have also apparently directly descended 

 from the Lithosiida'. The Syutomidie and Nyctemerida% which seem closely allied by larval 

 ■characters, have also directly descended from the Lithosiid;^. 



Finally, it appears that the Geometrid;e are a rather more primitive type, and have no 

 relationship to the ^SToctuidie, the latter having more or less directly descended from the Agaristida-, 

 the latter from the Uypsida- or an allied group. The fact that the young larvit of many Noctuidie 

 have onlj- two pairs of legs seems to have no phylogenetic signilicance. 



In this preliminary abstract space has prevented my giving details and figures to prove tlie 

 truth of the assertions and conclusions here presented. 



Hints on the orh/hi of the Noctitidw. — The Xoctuida' may have descended from the Agariotida^, 

 since the pupse^of several geuera I have examined are of the same type as those of Alypia and 

 Eudryas, having a similar lanceolate labium (second nnixillary paljji). It is possible that the 

 Agaristid* are the direct offshoots of the Uypsida^ or came from an extinct groui) closely allied 

 to them. ( )f this I can, from the want of specimens, only judge from the figures in Horsfield and 

 ^loore's Catalogue of Lepidopterous Insects, etc., Part II. The caterpillars of Hypsa, Eusemia, 

 particularly E. basalis, are not only much like ordinary Noctuiche, but are also closely similar to 

 those of Eudryas and Alypia, that of JE. basalis being humped on the eighth abdominal segment, 

 and with the dark bars and spots of the larvae of these Agaristids. Hence, (juite contrary to our 

 former prepossessions, it appears probable that the Noctuidic may be the descendants of th.e 

 Agaristid:e, iustead of being connected by the Deltoids with the Pyralids. That the N octuidie, 

 as well as the Geometriche, are a modern group is shown not only by the pupal and other 

 ■characters, but by the fact that they comprise so many closely allied genera and species, the pupse 

 as well as imagines possessing no vestigial charaeters. . 



The following tabular view will express in a tentative way my present views as to the 

 phylogeny of the Lepidoptera, or, in other words, the relatiouship of the suborders and of the 

 principal families, and will thus serve tcinporarily as a genealogical tree of the order. 



It will be seen by this scheme that the geuera of the Protolepidoptera, Palaeolepidoptera, 

 and those 2^eolepidoptera with incomplete pupre (including all the families up to Lithosiida?) have 

 narrow wings, the internal l)order of the wings, or "folded i)ortion" of Spuler. being slightly 

 developed. The flight of these genera is a tiuttering one and, in general, of short duration. 



In the later Lithosiida' we see a rapid enlargement of the folded portion or a widening of 

 the wings, and with this widening of the wings, with an increase in wing-power and the ability 

 to take longer and higher flights, we seem to have had a great increase in the nuinbei- of genera 

 and families, until in the butterflies, with their very broad wings, we have not only a fluttering 

 -and direct long flight, but also the power of soaring high in the air. On the other hand, the 

 S. Miss. 50 



