294 Mr. Voulion''s further notes upon the 



its tip is sometimes green. Before ecdysis the light red 

 tubercles ou the head of the last stage can be seen 

 through the skin of the fourth. 



Stage V. — A faint pink tinge is very common upon the 

 white stripes of this species, and it is especially noticeable 

 in this stage, although it also occurs in the fourth. It 

 occurs on the tubercles and ground colour, and is chietiy 

 developed on the 7th stripe, although it is often present 

 on the others, and even on the remnant of the subdorsal. 

 The two tubercles on the head are now blunt, but still 

 large : their summits are red at the beginning of the 

 stage, the bases yellow (see fig. 13, x 2, Plate YII.). 

 Very soon the red is entirely replaced by yellow (see 

 fig. 14, X 2. Plate VII.). The head is not generally 

 a bright blue, like the tail, but is greenish at first, 

 becoming greenish blue later (sometimes it is bright 

 blue). My larvae did not, as a rule, nibble off each 

 other's horns, but a few were treated in this way. I 

 found seven larvae on one small tree of Salix Babylonica, 

 and nearly all of them had lost more or less of the horn. 

 This injury may take place quite early in the life-history, 

 and I think that it is often extremely hurtful or even 

 fatal, as a considerable amount of fluid is lost. The 

 midrib-like appearance often occurs in this stage also, 

 produced by the dark dorsal line with a row of white 

 tubercles on each side. 



Conclusions. — I formerly suggested that the red spots 

 which sometimes appear in SnierintJius larvae are due to 

 reversion to a more brightl}^ coloured condition. It is 

 possible that the pink tinge to the white stripes is to be 

 explained in the same manner. But the brightly- 

 coloured and prominent tubercles on the head can only 

 be interpreted as of historic significance, indicating that 

 the shape and colour of this larva have become subdued 

 for protective purposes. The tubercles are important 

 both in shape and colour in the second and third stage ; 

 the colour becomes faint at the close of the fourth, and 

 in the last stage their shape has altered, becoming 

 inconspicuous, while the bright colour which now only 

 appears on the tips of the tubercles, disappears soon 

 after the beginning of this period. Thus we have a 

 character that strongly supports the interpretation of the 

 bright spots as due to reversion. The difference seems 

 to be that the history of the disappearance of the head 

 tubercles can be traced in every ontogeny, while the 



