( 170 ) 



of a long tongue-case in tin- short-tongaed Acherontia, lu-ovcs the short-tongue il 

 forms to be derivations I'lom h)iiger-tonguod species. There occur some peculiar 

 pupae in this subfamily : for instance, that of Langia, which is truncate apically 

 and has no cremastcr : tiie chrysalis of Cressonia with the preanal segments 

 flattened ventrally and carinate laterally ; the chrysalis of Phyllosphingia with 

 ])rominent tubercles serving locomotion. 



From the morpiiological characters of tlie imagines stated above! it will have 

 become clear that the genera Protamhuhjx, Aiiiplf/ptrriis, etc., represent the earlier 

 Ambulicinae, and are not separable from this subfamily. The elongate fore wings 

 are not a character confined to Protamhuli/j' and allies ; the comparatively strong and 

 long tongue is met with also among Poh/pti/chus and near relatives; and the lateral 

 " expansions " of the abdomen — they are tufts of scales — by wliich Hampson 

 cliaracterised the " ^/«4«^/c<««(s" are found only in the Indian genus Oxi/antbulyx. 

 On the other hand, tlie tongue is very short in some American forms which are 

 close allies of Am/jlgpferut>, and one of these has even the outer margin of the 

 forewing of the ordinary irregular type ( Troj/olcijnum). Callambidyx rubricosa, 

 hitherto considered a true Protambiih/x, is much more closely related to Sphinx ocellata 

 than to I'rotambidifx. The connection between the specialised species of Ambulicinae 

 and Protambuhix and allies is so close that there is no justification for separating 

 Protttmbnlyx and allies as another subfamily. This will more clearly be understood 

 by comjiaring the following key to the probable phylogenetic connection between the 

 various genera. It will be noticed that here, as among the Aclierontiinae (p. 4), 

 similarity in structure arrived at in consequence of the reduction or loss of organs does 

 not necessarily mean close relationship, the most specialised forms of the various 

 branches agreeing in several characters which they have independently acquired. 

 The various genera represent in this key steps in the development of the organs 

 mentioned ; each genns has besides, as a rule, its own specialisations not mentioned 

 in the key, making it obvious that probably not one of the genera is the ancestor 

 of the next higher one, but that both come from a common stem. 



There are 130 odd species known, which fall into 44 genera. The subfamily is 

 distributed all over the globe, except the far north. South and Central America 

 have few rejiresentatives, all of the lower type (ProUunhulyx') or closely related to 

 it. The headquarters of the subfamily are in Africa and India. On the former 

 continent many new forms will be discovered when the fauna, especially of the 

 western side, becomes more fully explored. 



Though the Ambulicinae do not fall into several definable tribes, there are 

 nevertheless some groups of genera closely related inter se. The four Neotropical 

 genera, Protambuli/x, Amph/pterus, Orecta and Trogolegniim, form one group ; the 

 last two genera are developments from ancestral Amph/pterus, having acquired 

 specialisations which are found again in other branches oi Ambulicinae: for instance, 

 the undulated distal margin of the forewing of Trogolegnum and the apical protibial 

 thorn of Orecta, specialisations accompanied by a strong reduction of the tongue, 

 and in the S by the loss of the friction-patch. Another rather well circumscribed 

 group are the eyed Ambulicinae, of which Callambulyx rubricosa is the most 

 generalised member, while Cressonia is the most specialised ; the group is Holarctic, 

 extending into the Oriental Region with two genera. 



The connection between Clanis, Pseudoclanis, and Platysphinx is very obvious, 

 and the relationship of Leucopklcbia and Leptoclanis with the same branch becomes 

 also evident on a close comjiarison of the structure, as pointed out wwAqx Lcptoclams. 



