416 BRITISH LEPIDOPTERA. 



the region of the oblique stripes and borders than elsewhere. In 

 M. atropos this is probably also the case, but the brightened ground 

 colour cannot, of course, be distinguished from the true oblique stripes 

 (which are also brightened ground colour). Hence what is described 

 above as a broadening of the oblique stripes may in reality be due to 

 their fusion with another of the elements of larval colour. The bright- 

 ness of the ground colour in the region of the oblique markings, due 

 in part to stripes and in part to the ground colour itself, produces an 

 effect which is very like that of S. ligustri. Below this . region the 

 ground colour becomes much darker and greener in both larvae and, 

 in both, the darkened colour spreads upwards in front of the borders 

 for the lower half of their length. In M. atropos, a fine white line runs 

 along the posterior margin of the inferior anterior half of each coloured 

 border, but it seems unlikely that this represents the true stripe. The 

 borders of M. atropos are violet for the chief part of their course, shading 

 into blue anteriorly and inferiorly, and also changing abruptly into the 

 same colour above, at the furrow which divides the two segments over 

 which each border extends. The upward blue continuation of the 

 border meets its fellow of the opposite side in a distinct V, of which 

 the apex lies close to the hinder limit of the posterior of the two 

 segments, traversed by the border. This backward prolongation of 

 the coloured border (as such) does not take place in 6 1 . ligustri, but 

 the lilac tint ceases abruptly at the furrow corresponding with that 

 which only divides the border of M. atropos into two rather differently 

 coloured halves. Nevertheless, careful examination will show that 

 the borders of .S. ligustri are carried back and form a series of Vs, 

 but only as darkened ground colour, and with the loss of all distinct- 

 ness. Hence each segment, from the ist to the 7th abdominal 

 (both inclusive), is crossed by parts of two oblique borders, of which 

 both the upper and lower parts are very conspicuous in M. atropos 

 while the upper in S. ligustri can only be detected after careful 

 examination. This is also true of the stripes themselves in S. ligustri, 

 while the stripe is altogether inconspicuous in M. atropos. The 8th 

 abdominal segment is only crossed by the upper part of one stripe 

 (the last) and its border, and here, also, the same contrast holds good 

 in the two larvae. In M. atropos, however, the border does not become 

 blue, but remains of the same purple as its anterior inferior continua- 

 tion. This border also becomes much narrower upon the 8th 

 abdominal segment after crossing the furrow which separates it from 

 the 7th segment. The larva of M. atropos is peculiar as compared 

 with that of 6". ligustri in the persistence of the "eighth stripe" in 

 the last stage ; but this latter larva possesses this feature in earlier 

 stages, and it can even be detected immediately after the last ecdysis. 

 In M. atropos, the " eighth stripe " is conspicuously represented by 

 the bright blue border which is present upon the ist abdominal 

 segment, and ends abruptly at its anterior margin, thus occupying an 

 entirely normal position. On the other hand, this is the only instance, 

 as far as- I am aware, of the " eighth stripe " gaining a coloured 

 border, the exception being doubtless explained by the fact that this 

 stripe consists of an upper part only (upon the ist abdominal 

 segment, and never extending downwards upon the 3rd thoracic 

 segment), and such upper parts possess coloured borders in the case of 

 M. atropos alone. The shagreen-dots form a very remarkable addition 

 to the markings of the adult over the dorsal surface of the abdominal 



