8o 
Winifred Smith 
naturally, and not to be eaten away by ants, for a transverse sec¬ 
tion shows a peripheral ring, 8 or 9 cells thick, inside the wood and 
bounded on the inside by continuous cell-walls. This cavity may 
perhaps be regarded as the result of adaptation to ant inhabitants 
on the part of the plant (see figures 10 and 2 ). 
Ingress is also facilitated by the plant. Grooves run vertically 
up the stem from the leaf axil to the next node, and these are pro¬ 
bably caused, as in Cecropia ( 2 ) by the pressure of the axillary bud 
on the soft young tissues and the longitudinal stretching of the 
stem. The grooves are somewhat wider at the upper end, but the 
stretching of the circumference has less effect than it has in 
Cecropia. The grooves are not present on all internodes, but in 
nearly all the grooves there are one, two, or three holes leading to 
the cavity within and placed near the end of the internode (fig 1 ). In 
some of the internodes without grooves there are holes which only 
penetrate a short distance and show signs of considerable cambial 
activity. Some of these holes are stopped by a plug of latex, 
possibly of caoutchouc. The latex cells, which are very prominent, 
have yellow, brown or red contents (see figs. 10 and 11 ). 
In young stems there is a distinct nodal plate, but this is per¬ 
forated in older material, probably by the ants, as in Cecropia ; so 
that there is free communication between the internodes and a hole 
is unnecessary in every one. 
As in Humboldtia laurifolia (r) and Duroia hirsuta ( 3 ) the food- 
bodies in Macaranga triloba are on the stipules. Here however 
they are found on the morphological under side, which becomes, in 
old retroverted stipules, the concave side where they are hidden 
from all but the initiated (see Figs. 1 and 3). They are of a very 
peculiar structure and consistency and much more work will be 
necessary before I can give a satisfactory account of them. They 
are not however unique among food-bodies eaten by ants in being 
abnormal, for Schimper says of the Beltian corpuscles ( 2 ): “Sie unter- 
scheiden sich jedoch von alien bekannten Driisen durch bestimmte 
Merkmale,.namlich durch bedeutendere Grosse, langerer 
Dauer, Reichthum an Eiweissstoffen, leichtes Abfallen beim 
Beruhren.” Their shape is spherical or pear-shaped, they are of a 
golden yellow colour, and their contents greatly obscure their 
structure. Absolute alcohol and clove oil cleared them somewhat 
and showed them to be multicellular, with external hexagonal 
marks, presumably cell-walls. They can be ruptured with consi¬ 
derable pressure under a cover-slip and then exude drops of a liquid 
