PROGRESS IN THE CLASSIFICATION OF THE SPHINGIDES. 65 



the genus il/ftr;-o//Zo.s.sfl of earlier writers, and Achenmtii)iae to Acherontia 

 of Herrich-Schaft'er. The old comprehensive family of Sphi)i(/idae 

 now became subdivided into Sphimjinae and CJiocrocaiiijtuuic, the 

 subdivision being made chiefly on larval characters, the anterior 

 segments of the larvje of Chocwcampinae being stated to be letractile 

 while those of the Sphinginae are not so. The 5th segment in 

 CJuierocanijiinae is also noted as abruptly broader than the remaining 

 segments. In the imago the head is said to be generally larger and 

 the tongue shorter. The old genus Siiwrinthus is divided into the 

 subfamilies Suieriuthinae and Ainbalicinae, the latter group being 

 erected to receive those species that, in the larval state, have the anterior 

 segment more tapering, and in the imaginal have slender antennae in 

 both sexes. Except for these two further divisions there was no 

 advance on the work of Herrich-Schatfer, in point of fact the only 

 real step was the separation of Clweroccmi [)inae from iiplniujinae, for I 

 doubt whether the Ambnlicinae are separated from Siiicrinthinae on 

 sufficiently important structural characters. The next important 

 writer on the Sphiimidae was Grote, who, in 1886, published the 

 Han-k-inot](» of North America. The principal step made was the 

 endeavour to show the relationship of the subfamilies inter se, and the 

 relationship of the SpJiiuj/idae to other families. The SnierintJiinae 

 are given a central (basal) position, and Grote expresses the belief that 

 the other groups may have arisen from its ancestors, while its present 

 members are perhaps nearer the Choerocinninnae than the Sphinf/iitae. 

 The SiiicrijUhiiuw show characters to be found in all the groups and 

 from this in part the above generalisation is deduced. The 

 Choerocanipinae are said to be related to the Sinerinthinae through 

 Amhidi/.v (i.e., AmhuUcinac of Butler), the Spliinijinae ihriowQh. the genus 

 Elloiia, and the Macroglosainae thvongh the genus Deidajiria. Speaking 

 of the relationship of the family Spliiwjidac, he discusses the anal 

 horn of the larva found in Bombijx mori and also the distension of the 

 anterior segments similar to that found in Choerocampinae. The 

 membranous tongue of the S)nerint/nnae is also urged as giving 

 support to the notion that the Bombijcidae are a nearly related family. 

 Because the Sphingid pupa has no silken cocoon, he considers this to 

 be a low or derivative character, and that the crepuscular or nocturnal 

 flight should be regarded in the same light. In conclusion the 

 Splnniiidcw are said to be evolved from the ancestors of the Cerato- 

 (■(Diipinae, a subfamily of the Bombijcidae. Following closely upon 

 Grote, Smith in 1888, in the Entomologia Americana, discussed 

 the relationship of the Sphiiujidae, and strongly insisted that no linear 

 arrangement of the genera could give an accurate idea of the true 

 relationships. Two main divisions of the family are given, the 

 Sjihini/ides with long corneous tongue and the Smerinthides to take all 

 other species. The SpJiinijidca are divided into two series by the shape 

 and proportions of the antennte, Macr()(/los,sinae forming one series and 

 Sphiui/ijioe the other. Of the relationship of the Sphingidae to other 

 families he says "I can find no close relationship with Sesia .... 

 Sesiids resemble Castniidae and Cossidae in so far as the larvje are 

 endophytes." The conclusions of Grote are adhered to in tracing 

 the relationship of the Sphimiidae with the Bombi/cidae through the 

 Smcrinthiiiae, 



The last work that we shall refer to is that done by Dr. Chapman 



