( 193 ) 



infcvantennal ovest. iSpnrs niie(|nal, short ones more than halt' tlio length of the 

 Ion;; ones, lon<j;er apical one of hindtibia more than hall' the length of the tirst tarsal 

 sefinient. Apex of forewinjr ncntuinate, not excised ; R^ of hindwinij in or below 

 centre of cell, D^ as long as, or longer than D^. Scales at lateral edi;e of eighth 

 tergite of c? prolonged to a triangular crest. 



Larva: head triangular ; wliite oblique side-bands; horn long. — Insufficiently 

 known. 



( 'hrysalis not sufficiently known : the fignres too vagne. 



/l/il). Indo-Anstralian Region, as far north as Japan, eastwards to the Solomon 

 Islands. 



Sixteen species. 



The head has a sharjily detined triinsverse baud between the antennae. The 

 band on the thorax does not occnj)y the whole of the distal two-thirds of the 

 inesothoracic tegula, the tip and sides remaining of the colour of tjie disc of 

 the thorax. The mid- and hindtibia bear a white basal spot as in Amplypterus. 

 The submarginal line of the forewing begins at the very tip of the wing : the line 

 in tlie apex of the cell is more or less oblique, often horizontal. The hindwing 

 lias a band beyond cell and one or two dentate lines in outer half; the distal line 

 is often absent or vestigial. The jiatch of modified scales on the clasper is more 

 or less ovate. 



The species fall into three groups: siihstrigilis-gvon\<i (12 species), canescens- 



gronp (one species), semifen-e7>s-gTo\vp (three sjiecies). The har[ie is of a different 



type in the three groups. The species of the first group bear all a rather close 



resemblance to each other, and are easily confounded. They required a closer 



examination and more careful study than had hitherto been bestowed upon them. 



The differences. in jiattern, as pointed nut in the key here appended, are corroborated 



by differences in strnctnre in both sexes, so far as both sexes are known to us. 



No notice has been taken of the great dissimilarity between the larvae o'llituruta 



and mihittriqUifs liy those authors who put the whole lot of species allied to snbMiic/ilis 



together under one name. The statement made here and there that sirbsfru/ilis h 



very variable individually is based npon the assumjition that the different individuals 



called KubstrigiUs belonged all to the same species, which assumption is quite 



incorrect. In Fainm lirit. Iml., Moth& there are six species mixed together under 



»Hhstri<iHi!<. They are not mere geographical forius, but species of the same conntry, 



constantly differing in colour and structure. It was not carelessness which misled 



Harapson to treat so many distinct species as synonymous, and misled also one of 



the authors of the present jiaper to suggest that aur/pr/ai/s, sericeiprniiig, etc., were 



snbRpecies or ]ierhaps only aberrations of one species, but the reason underlying 



those opiniiins was of quite a different kind. Bntler, and also Moore, had described 



as distinct in several instances either the sexes or geographical races (snbspeciesj. 



When this was noticed, reaction carried me (Rothschild) too far, and I nnited also 



forms which are really distinct, because I found that the characters by which the 



respective authors separated the species did not liold good. Though I was right in 



I this latter respect, I was wrong in assuming, with Hampson and others, that there 



were no constant differences, the actual distinguishing characters of the various 



•pecies not having been known until we studied the insects for the jiuriiose of the 



[ireHCht Revision. 



Most of the si)eeics are conlincil to the Himalayas ; their range will doubtless 

 be fonml to be less restricted than it ajipears to be at jiresent, when the Sphingid 







