ADDENDA. 287 



We never found either larva or pupa of this butterfly without an attendant guard 

 of the ferocious red, or yellow, tree ant, (Ecophylla smaragdina, as we believe the fiend 

 is called. Half-a-dozen of them, kept with the larva in a bottle, lived quite happilv 

 for a week, and seemed to require no food besides what they obtained from it and from 

 the leaves with which it was supplied. The larva was in no degree dependent on them, 

 as we proved by rearing some from a very small size without ants. From this it would 

 appear that the advantage which the larva derives from the alliance is protection 

 against enemies. The house swarmed with a small brown ant, very active in killing 

 and carrying off sickly larvae, or butterflies just emerging from the pupa. We 

 introduced some of these to a centaurus, and immediately they took it into their care 

 and showed that they thoroughly understood the management of it. First the larva 

 was assiduously caressed all over, especially the head ; then the ant went to the 

 11th segment, touched the gland gently with its antennse, and was rewarded with a 

 drop of honey, which it licked up at ouce. This process was repeated many times. 

 Now, supposing these ants to be nine years old, like some of Sir John Lubbock's, it is 

 not possible they ever could have seen a centaurus larva before, for the house in the 

 walls of which they had their nest stood actually on the sand of the sea-beach, a mile 

 from the nearest spot on which we ever saw this butterfly, or found its larva. How did 

 they understand it so well ? Perhaps they had had some practice with Lampides 

 i£lianus, the larvae of which we sometimes found in the garden. 



ARHOPALA AMANTES, Hewitson (p. 150). 

 Plate 705, A, fig. 5, larva. 



Larva. — Feeds on the same plants, is very similar, but differs in the hairs being- 

 longer. It is a much paler catej-pillar, the 4th and 5th segments and the 10th having 

 the back a clear chrome-yellow instead of brown. It is usually attended by masses of 

 red ants. 



Pupa. — So similar, that difi"erences can only be discovered by microscopic 

 examination. 



ERRATA. 



p. 31. Plate 648, figs, lb, c ; should be lb, ^ , Ic, ? . 



p. 117. Simiskina. Foreidng with three sub-costal nervules, not four. 



p. 133. Erase last two lines ; should be, " We give Horsfield's and Moore's figures, and de Niceville's 



descriptions made out from those figures." 

 p. 194. dodonwa should be dodonxa. 

 Plate 662, figs. 2, 2a, 2b. The artist has left out the tails ; they should be similar to those of 



Chrysophamis susanus. 



