Ridley.—Symbiosis of Ants and Plants. 481 
Macaranga , but they usually fall off after they have performed their 
function of protecting the bud. In M. triloba they not only remain 
attached, but reflex themselves and continue to grow. The bladder-glands 
occur on other species than the myrmecophilous ones, and perhaps perform 
an excretory function. 
In the myrmecophilous species they continue to develop or rather 
to increase in size and to produce an oily liquid in their interior, which 
is attractive to ants. Only a very few of these bladder-glands develop 
into food-bodies. Indeed their great numbers would make it impossible 
for all so to develop in so small a space. Why some should so develop 
and others not is by no means clear. 
Hollow stems are not by any means common in plants in this region, 
and no Macarangas which are not otherwise adapted for myrmecophily 
have any part of the stem hollowed, so far as I have seen. In the case 
of Clerodendron fistnlosum , Beccari suggests that the inflated stems produc¬ 
ing a strengthening of the plant allow it to grow straight and strong above 
its competitors, and that the parents of the plant appear to have been 
slender and herbaceous and badly adapted to sustain the lot of its existence 
among tropical vegetation, mostly woody. In the view of the relationships 
of these Macarangas to the solid-stemmed non-myrmecophilous species, 
and the occurrence of many solid-stemmed Clerodendrons in our forests, 
I should doubt the validity of the suggestion. 
It is clear, I think, that the advantage which the plant derives from the 
presence of the ants is protection from the attacks of caterpillars and 
possibly thrips. It may be taken as certain that the greatest time of 
danger to the plant from insect attacks is while it is in the seedling stage, 
while it possesses only one growing-point. The destruction of the terminal 
bud may cause the death of the whole plant, and even if the plant has 
already branched, which it does not usually do till it is at least ten feet 
tall, a caterpillar can and doubtless would destroy the second bud after 
finishing the first. Buds of plants in the tropics are usually very carefully 
protected by coverings of bud-bracts, wax, fur, or by other methods. Even 
in full-grown trees the destruction of the young leaf by insects or fungi does 
more harm to the tree than an extensive raid of caterpillars on the adult 
foliage. The damage caused to a tree by the destruction of the adult 
foliage by caterpillars seldom in a natural state causes any very serious 
permanent injury to the tree, but the continued damaging of the just 
expanded leaf is most injurious. 
In an adult Macaranga the ants only frequent the ends of the boughs, 
moving on from internode to internode as they develop. The stiffly 
coriaceous adult leaves of M. hypoleuca , and the firm but less coriaceous 
ones of M. triloba , are not attacked by caterpillars, and do not require 
protection. The defence required is only that of the young leaf during the 
