Ridley.—Symbiosis of Ants and Plants. 
47i 
include species so very different in habit and structure it would probably 
be more satisfactory to break it up. 
In studying the Malayan species I find that a certain number are 
myrmecophilous, and as three of these species are abundant in the Botanic 
Gardens of Singapore, I have for some time been examining these plants, 
and making notes on them with a view of finding out how far they were 
actually modified so as to attract and ensure a regular symbiosis, and I 
think that the following observations will show that the Macarangas of this 
group do fill all requirements for their classification as truly myrmecophilous. 
I have also examined the herbarium specimens in the herbarium of the 
Botanic Gardens, Singapore, in order to compare the myrmecophilous 
species of the genus as at present understood with those which are not 
tenanted by ants, and which possess no organs modified for myrmecophily. 
In all the myrmecophilous species the stem in the young plant is hollowed 
by the disappearance of the pith, and in the hollow stem live the ants, which 
obtain access to the outside of the plant by holes perforated through the 
side of the stem at irregular intervals, but one or more appears in each 
internode. The ants also perforate the septa of the nodes so that a 
communication exists between all the internodes. In adult trees the ants 
usually only reside in the ends of the branches, the hollow in the trunk, 
though long persisting, being abandoned. 
In most of the species of the genus, the buds are protected before 
opening by an opposite pair of triangular or lanceolate sessile leaves or bud- 
bracts, which have been called stipules in various papers, but which seem 
to me to bear no relation to the stipules of the Leguminosae nor of the 
Rubiaceae, but are rather of the nature of bud-scales, and I shall call them 
bud-bracts throughout this paper. A certain set of the plants included 
under the genus Macaranga in the flora of British India and elsewhere do 
not appear to possess these bracts at all; such are M. trichocarpa , Muell., 
M. carolinensis , Volkeres, M. cumingi , Muell., M. javanica , Muell. None of 
these are myrmecophilous. Another series possesses well-developed bracts 
which are persistent for some time after the bud has expanded, but are not 
in any way modified for the use of the ants, nor is the stem hollowed or 
occupied by ants. These trees in fact are not myrmecophilous. Such are:— 
Malay Peninsula. 
M. tanarius , Muell. Arg. 
M. Curtisii , Hook. 
M. popidifolia , Muell. Arg. 
M . depressa , Muell. Arg. 
M. bicolor , Muell. Arg. 
M. megalophylla ) Muell. Arg. 
M. riparia , Engl. 
Borneo. 
Philippines. 
Malay Peninsula. 
New Guinea. 
The bracts in these persist for so long that there are often as many as 
