sphingin^e. 267 



the oblique stripes have reached a higher development, jemg always 

 of two, and sometimes of three, colours (e.g., drupifexxrum — white, 

 red, and black), whilst in the species of Smerinthns they only 

 occasionally possess uniformly-coloured borders. 



We have already dealt with the larval structure of the Sphingids 

 (sens, restr.) (taking Sphinx ligustri and Agrius convolvuli as types) 

 (anted, vol. iii., pp. 367-368). The only point to repeat is that, in the 

 1 st instar, the primary tubercles agree in their arrangement with those 

 of the Sesiids and Amorphids, i.e., tubercles i and ii on the abdominal 

 segments form the typical trapezoidals, but on the meso- and meta- 

 thorax they arise from a single plate or raised area, whilst iii forms 

 the typical supraspiracular, iv is subspiracular, and there is a pre- 

 spiracular (which is said by the authorities to be v, moved forwards 

 and upwards) quite at the front of the segment, vi being 

 nearly at the base of the proleg *. One of the most interesting 

 features of the newly-hatched larva is the absence of shagreen- 

 tubercles and secondary hairs, which appear suddenly at the 1st 

 moult, and are retained throughout larval life, although they gradually 

 atrophy, but yet can still be traced in the last stadium in the species 

 examined. The structure of the caudal horn is interesting from 

 the fact that it is thickly clothed with short bifurcate hairs, whilst 

 the tip carries two setae corresponding with those arising from 

 tubercles i on the other segments, these setae being somewhat 

 enlarged, funnel-like apically, and with traces of spiculation. The 

 caudal horn has no doubt originated as a raised area, bearing 

 the two tubercles i, which has gradually developed and risen until 

 it has assumed the character of a horn, always carrying on its 

 summit the two tubercles, distinct but varying in position, probably 

 at first with something of the double character now seen in the 

 horns of the larva of Lophopteryx canielina, coalescing, however, and 

 forming a common base long before being so well-developed (sepa- 

 rately) as that of the latter species. In the larva of Agrius convolvuli, 

 Poulton notes (Trans. Ent. Soc. Lond., 1888, p. 519) that the bifidity 

 of the horn varies immensely in the 1st larval instar, is somewhat 

 shorter than, but is covered with hairs somewhat similar to those 

 on, that of Sphinx ligustri. There appears to be no doubt that the 

 bifurcation of the horn, which is common throughout a great many 

 very different sections of the Sphingids (sens. lat.J, is quite a 

 primitive and ancestral feature. Poulton connects (loc. cit., 1885, 

 p. 303) the use of the horn with the longitudinal markings ex- 

 hibited, and states that the horn in Sphinx ligustri and Smerinthus 

 ocellata exhibits movements synchronous with the contractions of 

 the dorsal vessel. The development of the typical longitudinal 

 lines and oblique stripes during the growth of the larva in the 

 1 st instar and before the 1st moult may also be mentioned. 



Among the lower Sphinginae the larval structure offers consider- 



* The hopelessness of using some of Poulton's early work is amply demon- 

 strated by comparing the above description with his figure of the newly-hatched 

 larva of Sphinx ligustri (Trans. Ent. Soc. Lond., 1885, pi. vii., figs. 1-2); on the other 

 hand, his drawing of the larva of Agrius convolvuli floe, cit., 1888, pi. xv., fig. 2) is 

 very fairly accurate, except possibly in the dorsal thoracic setae, that do not show 

 the arrangement of rising from a common base common to the higher Sphinginae^ 

 whilst vi is missed altogether. 



