— 273 - 



In old galls generally the breathinghole of the larva is open, seldom 

 closed by a yellow coloured stuff; in some cases the membrane is pierced by 

 small holes of different size, evidently not made by the legal inhabitant but 

 by its parasites. 



The surface of old galls is entirely smooth. I never found them on the 

 mid rib of the leaves and always on the leaf surface. By cutting the galls 

 longitudinally they show a curved larval chamber as shown on plate XI, fig. 5. 



As shown on plate IX and X the surrounding normal pandan leaves are 

 much disfigured by the pressure effected by the galls, as the latter are much 

 thicker than the normal leaf, as can be seen on plate XI fig. 4. 



The development of the galls (plate VIII, a — f). 



Old galls appeared nearly always to be deserted by its inhabitants, in 

 some rare cases they contained parasites. As galls are as a rule formed in still 

 growing, not yet quite developed parts of plants, it was clear that the young 

 not yet quite developed galls must be sought for in the younger and youngest 

 leaves in the heart of the plants and indeed they were found there in all 

 phases of development (see plate VIII). As the hidden basal parts of young 

 leaves are not yet green but yellowish white, the young galls are coloured as 

 the young leaves. When the leaves grow out while the galls develop and 

 the latter therefore are exposed to light, the galls turn pale green. 



In this phase of development the larva has already closed its window 

 by a sheet of silk. 



As the larva is only gradually enclosed by the growing tissue, it is very 

 probable, that the egg (or the larva) is laid on the epidermis and not in 

 the leaf, otherwise the very young larva should have been found within the 

 leaf tissue. 



The egg was not found. The smallest gall which could be found was 

 nothing more than a very faint swelling or blister in the leaf tissue, measuring 

 2 m.M, in length. In the centre a very shallow excavation could be observed 

 wherein an exeedingly small nearly translucent larva with cloudy orange 

 spots was feeding. 



Bigger older galls show a deeper and larger excavation (plate VIII fig. 

 a — c) and a surrounding wall or edge of greater height. In still older galls 

 of 1 — 1J4 c.M. length the walls of the excavation are drawing near each 

 other on one pole (fig. d) leaving at the other pole a hole; and still later 

 the walls of the narrow excavation are meeting entirely only leaving the 



