marlhifis and attitudes oj lepidnpterous lame. 305 



otherwise complete series must tend to attract attention. 

 Such a break is rendered necessary by very deep-seated 

 anatomical facts in any series of markings that arise 

 round the spiracles, because of the absence of these 

 latter upon the meso- and metathorax. The series 

 could be rendered continuous in either of the following 

 ways :— (1) by concealing the spiracle on the prothorax 

 (which is the real difficulty, for the series is complete, 

 and ends at the 1st abdominal without it) ; (2) by 

 acquiring a different mark round this spiracle so as to 

 prevent it from forming part of the series; or (3) by 

 continuing the marks upon the meso- and metathorax 

 regardless of the absence of spiracles. This latter is 

 seen in the extreme forms of the spotted varieties of 

 S. popidi. The second is seen in M. stellatarum, where 

 the anterior spiracle is red (and far less conspicuous) ; 

 the others black. The very interesting cases of S. ocel- 

 latus and jwjmli are instances of the first method. In 

 the protective attitude the anterior spiracle is completely 

 concealed by approximated folds of skin. In S. impidi 

 also it is less conspicuous at all times. This is very 

 interesting, because the spiracles are brightly coloured 

 (in both species), and that upon the prothorax would be 

 far_ more conspicuous than the others in the Sphinx 

 attitude. In S. lif/ustri the spiracles are far less con- 

 spicuous, and the interruption does not attract attention. 

 There are probably many other instances of the different 

 methods by which a series of spiracular markings may 

 be rendered continuous. In many Noetua larva) the first 

 and last spiracles are much larger than the others, but 

 this is probably due to physiological needs, and the 

 series is not rendered conspicuous by size or colour. 



12. Phytophagic coloration.— (1) S. oceUattis. —Bming 

 the past summer I bred a number of larvae from the egg, 

 and I certainly found some considerable difference pro- 

 duced by the_ different food-plants. Thus apple and crab 

 caused whitish-green larva; while Salix cincrea, and 

 especially S. rubra, produced adult larvae tending 

 towards the yellowish form. S. viminalis did not pro- 

 duce nearly such white larvse as the apple, but, con- 

 trary to my expectation and experience, the larva were 

 almost intermediate. Thirty larvae were experimented 

 upon, in five batches, fed respectively upon the food- 

 plants mentioned above. All the larvae, except two or 



