346 BRITISH LEPIDOPTERA. 



versely, those with the larger amounts hatch comparatively early, 

 but are compelled to continue the process outside the egg. Hence 

 the length of the first stage would be correspondingly greater in 

 the latter case, but no effect would be produced upon the later 

 stages. It is probable that the length of larval life (exhibited by 

 these particular individuals) was unusually great, and that the 

 larva becomes mature at an earlier period in the localities in which 

 it breeds regularly. The closely allied S. ligustri has a larval 

 period of about five weeks, and the same is true of Smerinthus 

 ocellatus. In this country, the leaves of the foodplant are extensively 

 killed by frost during a period which corresponds with the last part 

 of the larval life — a time when growth is especially rapid. I could 

 only get food by obtaining it here and there in many places, and 

 my larvae would certainly have perished of starvation in the wild 

 state. In addition to this, the larval life would have been somewhat 

 longer in the latter state, for the larvae were kept indoors at a com- 

 paratively high temperature. They were fed so carefully throughout, 

 and the food was kept so fresh, that there was no protraction of 

 life owing to the causes which frequently operate upon larvae in 

 captivity. Temperature has a very great influence upon the length 

 of larval life, as I have especially observed in the case of Smerinthus 

 ocellatus during the past cold summer (1888). It, therefore, seems 

 probable that this larva pupates comparatively rarely in this country, 

 and that, in warmer countries, the larval life is shorter, approaching 

 that of Sphinx ligustri!'' 



Larva. — * First stadium (newly-hatched) : The body very pale 

 yellow, with a slight greenish tinge on the anterior part of some 

 larvae ; the head of a very pale, brownish-yellow colour and of 

 rounded shape. (Hellins describes the larva as at first white.) 

 The caudal horn, which from the first is held very nearly upright, 

 is quite pale and transparent immediately after leaving the egg, 

 but it very soon darkens and becomes black. [It is very surprising 

 that the deep colouring-matter of the ova should produce so little 

 effect upon these pale and colourless larvae.] The newly-hatched 

 larva when extended is 3*7 5mm. in length, and there is very little 

 individual variation. [The newly-hatched larvae of S. ligustri were 

 found to be 5*2511:11x1. in length.]. Shortly after the first meal 

 the larvae became green, and, by September 12th, they had acquired 

 a glistening appearance quite unlike anything I had hitherto seen in 

 the larvae of Sphingidce. On September 13th, microscopic examination 

 of the two oldest larvae showed that the colour had extended to the 

 blood, which was of a pale yellowish-green colour in the claspers 

 and other parts remote from the digestive tract. The lateral long- 

 itudinal tracheal vessel can be distinctly seen with a lens, upon 

 each side of the transparent larva, and the dorsal vessel . and 

 Malpighian tubules are also plainly visible. The dorsal vessel 

 forms a very distinct feature throughout the stage. The fact that 

 the green colour of the larvae is, at this stage, chiefly depen dent upon 



* If we had to criticise [antea, pp. 307-310) adversely Poulton's faulty descrip- 

 tion of the 1st stadium of the larva of Sphinx ligustri, it is only fair to say that, 

 although no structural details are given, his account of the development of the 

 markings of the larva: of Agrius convolvuli has no equal for completeness in the 

 description of any other Sphingid species known to us. Without his work, this section 

 of our life-history of A. convolvuli would have been practically blank. 



