ME.MOIKS OF THE NATIONAL ACADEMY OF SCIENCES. 35 



"The most natural iirrangement would be for the Sphingida> to form the end of one special 

 line of Boinl^yees, the order being the exact reversal of that given above" (p. 573). 



In an essay" published in 1890, I indorsed Poulton's conclusions, remarking that while the 

 Sphingidffi had probably descended from forms like the more generalized Ceratocampida;, there 

 were some points in the iinaginal characters which appeared to forbid the idea that ihey have 

 immediately descended from Aglia. It now appears that this genus does not stand alone, but is 

 closely related to Arsenura, etc., the group Agliinw being a South American one, with a single 

 Eurasian genus in the ArctogaMc realm. I may be permitted to quote the view then presented. 



"But the origin of the Sphingidw from forms like our modern Ceratocampid* is supported 

 >)y a fact not mentioned by other observers, i. e., the similarity in shape and the great size of the 

 anal legs of Sphingidie and those of the Ceratocampida?. 



"Anyone who will compare the larvae of the two groups will be struck with the resemblance. 

 The sphinx-like attitude is also assumed by Eaden Imperkdl^ while feeding, and, taking together 

 this identity in attitude, the presence of a caudal horn and the general shape of the body, I do 

 not see wh^- the Ceratocampid;^ may not be regarded as an archaic group f i-om which the Sphing- 

 ida:; may have sprung, while the former may have originated from spined Notodontian larvsv, 

 such as CEdemasia conclmni, the Notodontians being apparently the most generalized forms of 

 all the Bombyces. and also as regards the larvte, being the most plastic forms; either assuming 

 the greatest variety of ornamentation, or being quite unadorned." 



In his excellent monograph of the Sphingida? of America north of Mexico'" Prof. J. B. Smith 

 divides the family into four groups or subfamilies: Macroglossinse, Cha^rocampina?, Sphinginre, 

 and Smerinthina?, in the descending order, regarding with others the Macroglossinffi as the most 

 specialized group, and the Smerinthina' as standing at the foot of the series, having a "small 

 retractile head and obsolete tongue."' He considers them as " insects thoroughlv bombyciform 

 in habit and appearance, but completely sphingiform in larval and imaginal character." He also 

 briefly suggests more clearly than an}' previous author, though not in a detailed way. the 

 resemblance of the Smerinthina? to what he calls the Saturniida\ stating that the group Sme- 

 rinthinte "seems to tind closer allies in the Saturniida' through Cressonia to the most typical 

 Smerinthina^." 



I have, after a somewhat prolonged study of the Ceratocampida?, compared them with the. 

 genera Cressonia, Marumba, and Paonias, and have been greatly interested and surprised to find so 

 many vestigial ceratocampid characters in the larva, pupa, and imago of the Suierinthinie. The 

 result is to prove, at least to my own satisfaction, that the caudal horn is only one of a numlier 

 of characters which indicate the direct descent of the Sphingidw from the Ceratocampida". and 

 most probably from the most primitive subfamily, the Ceratocampina?. 



The two diagnostic characters which separate the more primitive and generalized SphingidiB 

 from the Ceratocampida? are the position of the tubercle of the spiracular series, or e of Dvar. in 

 the larva, and the presence of an additional vein (III 2, ladius 2) in the forewings of the imago. 

 As stated further on, these appear to be sudden acquisitions which originated during the period 

 when the group diverged from the parent ceratocampid stock. It should be observed that the 

 tubercle r is of the same shape and structure, the difference between the larvw of the two families 

 being in regard to its position. 



• Larval featurvs. — We will now, beginning with the larval cliaracters. give the g'-ounds for 

 our opinion that the Sphingid;^ have directly- descended from the Ceratocampinw. 



The young larva (Stage I, PI. XLII) of (Watom'm ami/ntoi\ in the shape of the head and pro- 

 portions of the body, the shape of the suranal plate and anal legs, is the same as in the young 

 of EacJes ittqyerlalh. I can see no distinctive family characters in the parts of the head and organs 

 of mastication, in the shape of the two divisions of the clypeus; that of the cleft labrum and of 

 the antenna^ are nearly identical in Eades and Ceratomia. It is doubtful whether there are diag- 

 nostic primary family features in the head and mouth parts of lepidopterous larvse in general, 



"Notes on some points in the external structure and i)hylageny of lepidopterous larvse. Proc. Boston Society 

 Nat. Hist. XXV, May. 1890, p. 100. 

 6 Trans. Amer. Ent. See. XV, 1888. 



