472 Ridley.—Symbiosis of Ants and Plants. 
six pairs in M. popidifolia on the end of a branch representing the develop¬ 
ment of as many internodes, often the lowest reflexed from its bud. In 
M. megalophylla these bracts are very large. The myrmecophilous species 
fall into three groups. In all the bracts are long-persistent as in the last 
group, and the stem or branches are hollow and inhabited by complete nests 
of ants belonging to the genus Crematogaster . Nearly if not all the species 
have wax glands on the young leaves, and the young parts of the stem are 
coated with wax, and on the margins of the young leaves red cylindrical 
nectaries are borne on the ends of the nerves. These drop off as the leaf 
attains its full development, leaving no trace. The leaves in all are peltate, 
entire, or three-lobed. 
The three groups into which these fall, or perhaps it would be better 
to say the three modifications which we find, are as follows:— 
(1) M. sp. from Sarawak in Borneo, where it was collected both by 
Mr. Hullet and Dr. Haviland. In this plant the bracts are very large, 
lanceolate, acuminate, deflexed, coriaceous, not appressed to the stem, but 
concave, so that they may provide a nidus or feeding ground for ants. 
It is a very remarkable plant and quite unlike any other known to me, but, 
unfortunately, I have not seen it alive. 
(2) The second group is represented by M. hypoleuca , Muell., of the 
Malay Peninsula and M. caladifolia , Becc., of Borneo. 
In this group the bracts are lanceolate erect and do not bear food- 
bodies, but food-bodies are borne on the backs of the young leaves before 
they are expanded. 
(3) The third group is the most elaborately myrmecophilous group of all. 
It includes M. triloba , Muell., M. Griffithiana , Muell., M. Hullettii , King, 
and, I believe, also M. Hosei> King, but I have not had an opportunity of 
examining this latter alive, and my herbarium specimens are not very good. 
In these the bud-bracts after the expansion of the bud are reflexed and 
continue to grow, till they come in contact with the stem, to which they are 
so tightly appressed by their tips that they form a ringlike body almost 
completely surrounding the stem, and so concave beneath, that the ants, 
which find their way in at an open corner near the base, can not only hide 
within, but occasionally bring their larvae there. The under side of these 
bracts bears many white globular or more or less pear-shaped bodies (the 
food-bodies), which are at first attached to the lower epidermis of the bract, 
but later become detached and are conveyed by the ants to the nest in the 
hollow stem, where the larvae are fed on them. These food-bodies are not, 
so far as I can make out, commonly, if at all, borne on the under surface of 
the young foliage leaf as they are in M. hypoleuca , although wax glands 
occur abundantly there. They seem to be continuously produced on the 
under side of the bract, as they are to be seen of all sizes, and both attached 
and free. They much resemble the figure of the 'food-bodies ’ of Cecropia 
