AGRIUS CONVOLVULI. 



35? 



marises the results of the comparison made above. The real 

 break is between the three green larvae and all the rest, and the 

 differences between the former far outweigh those between the 

 latter, Hence the following classification expresses the relation 

 between the varieties : 





GREEN VARIETIES. 





DARK VARIETIES. 





Degrees of Colour 



I 



2 



3 



4 



4a 



\b 



¥ 



Afd 



¥ 



Division I of the last stage 



(I)* 



— 



— 



(2)* 



(3) 



(4) 



(5) 



(6) 





Division II of the last stage 



— 



(I) 



— 



— 



— 



(2) 



(3) 



— 



4* 



Division III of the last stage 



— 



— 



(i)* 



— 



— 



— 



(2) 



(3)* (4) 



(5) 



The asterisks indicate the larvae which are figured on pi. xvi. It 

 is very curious that each of the three divisions should have produced 

 a single green variety. In this stage, for the first time, the shade 

 of the larvae followed, upon the whole, the arrangement of the 

 last stage, although the parallelism was far from being complete, 

 as we see in the appearance of a green variety in the darkest 

 division (III). The bright orange spiracles form a very prominent 

 feature in this stage, together with the bright yellow longitudinal 

 stripes in some of the dark varieties (see fig. 4). At the close 

 of this stage, during the resting-period, the prothoracic dorsal 

 plate became very conspicuous, the cuticle around it being thrown 

 into wrinkles from the strain to which it was exposed, while the 

 denser and thicker plate was able to resist the strain. Its appearance 

 is shown in pi. xvi., fig. 11 (x 7). The plate, being darker in colour 

 and without wrinkles, is very prominent against the surrounding 

 cuticle. The 6th larva of division III was mature in the 3rd stadium, 

 on October 8th; on October Toth it had entered the resting-period ; 

 on October 12th (4.30 p.m.) it had just changed its skin. The 

 larva was that known as 8 in the division of "10 pale larvae of 

 the 2nd stage" (see comparison at end of 3rd stage). It is represented, 

 in the 3rd stage, on pi. xv., fig. 12. This larva died on October 16th ; it 

 was apparently a very dark variety, but was not sufficiently advanced 

 in the 4th stage for a safe comparison with the other dark larvae. 

 Fifth stadium : All the larvae became dark varieties in this stage. 

 Immediately after the change of skin the colour had not darkened 

 into its permanent shade. Thus the head was yellow and green. 

 The dorsal prothoracic plate, and the anal flap and horn were the 

 lightest parts of the larva immediately after ecdysis. It is probable 

 that this is due to the protection afforded by the thickness or density 

 of the cuticle over these parts, including the head. The darkening 

 of the larva being due to the action of the air, it is probable that 

 these parts alone would be completely protected from it until after 

 ecdysis had taken place. The shagreen-tubercles of the previous 

 stage are very distinct before the colours have darkened, appearing 

 as white circular areas, surrounded by a rather deeper shade of 

 ground colour, each containing a dark point in the centre. (This 



