APPENDIX. 485 



The neck collar is usually light grey sometimes tinged with 

 pink. More rarely it is dark grey. I have noticed this phase 

 of development most in very late specimens (Powell). 



[Page 251.] Egglaying. — Always laid singly, on upperside 

 of leaf, low down on plants. In spite of the comparative abundance 

 of young larvae in September, 1903, at Hyeres, close search for ova 

 only resulted in the finding of one empty eggshell which was on the 

 upperside of a small leaf on the growing point of a branch low down 

 on the plant; the very young larvae are nearly always found in a 

 similar situation (Powell). 



[Page 251.] Ovum. — Rather pale green, smooth and shiny; 

 almost round in outline, but with a flattened base. The empty 

 shell is practically colourless, thin, and nearly transparent (Powell). 



[Page 251.] Habits of larva. — The very young larvae are 

 nearly always to be found on the small leaves of the lower branches 

 of a plant, whilst, in their last stage, the larvae are generally found at 

 the top of a shrub feeding without a pause at the extremities of the 

 juicy shoots, and frequently even eating the flowers. I am sorry 

 I have no exact data as to the duration of each larval stage, but 

 growth is rapid, the larval state lasting about three weeks or a 

 month in the warm weather of late summer and early autumn 

 (temperature on an average=max. 25 C, min. 12° C), but the 

 individuals found late in October are retarded by colder weather 

 and do not attain a large size. As just noted, the young larvae 

 are to be found mainly on the undersides of the leaves, always near 

 the top of a tender shoot, and low down on the shrub or tree. They 

 move from one branch to another occasionally, without having 

 finished all the tender leaves on the last. As soon as the 5th instar 

 is reached the larvae make for the top branches of the shrub and eat 

 the freshest leaves obtainable. The larva feeds continually in this 

 stage, night and day, only ceasing in order to reach another branch 

 when all the tender leaves (and some of the tougher ones) on the 

 branch it has just left, have been entirely or partly eaten. It is easy 

 to detect the presence of a fullfed larva on a tree by the appearance 

 of the tops of the main branches. If a larva happens to meet with 

 a flower or flower-bud it will not eat leaves again if flowers are 

 available, and, as the trees are in blossom at the time the larvae are 

 feeding, this often occurs. This makes them rather more difficult to find, 

 for it is not easy to discover their traces amongst the masses of flowers, 

 however, by searching on the ground beneath the oleander bushes, 

 large droppings are seen at once should a larva be present. The 

 oleanders cultivated here, at Hyeres, have white, pink, dark pink or 

 salmon-coloured flowers, and larvae are to be found on all of them, but 

 rather more commonly on the white-flowered trees. The leaves of the 

 white oleander are generally more tender than those ot the other 

 kinds, which may account for the preference. Wild oleanders, with 

 pink flowers, grow in some of the dry watercourses to the east of 

 Hyeres, and the leaves of these wild shrubs are tougher than those of 

 the cultivated specimens. One autumn, when larvae of D. ?ierii 

 were plentiful in the gardens and oleander hedges near Hyeres, T 

 examined numbers of these wild trees, but found no trace of larvae 

 of this species on them. When at rest, the larva raises its head and 

 forelegs, contracting its thoracic segments, so that the great blue and 



