DISCHIDIA WITH DOUBLE PITCHERS. 385 
nesting-material ; (2) that the pitchers have no other adequate 
source of supply. To what extent the welfare of the plant is 
dependent upon the food-materials obtained by these pitcher- 
roots is unknown. Their abundant development and their 
behaviour towards the soil in the pitchers, point to their being of 
no inconsiderable importance. These four species of Dischidia 
must therefore be regarded as myrmecophilous to a marked 
degree. 
The presence of ants in the pitchers being highly desirable 
from the point of view of the plant, we are led to enquire what 
special characters, likely to be attractive to ants, the pitchers 
have acquired. In the first place, they are convenient shelters 
and nesting-places, and it may be presumed that the narrower 
entrance and more commodious form renders them more suitable 
for these purposes than the simpler pitchers of D. Rufflesiana. 
The danger of drowning which is risked by ants nesting in the 
pitchers of D. Rafflesiana must be considerable, though drowned 
ants are stated to be rarely found*. The branching root-system 
entering through a comparatively large opening (4mm. in diameter 
in adult pitchers grown at Kew) would seem to provide them with 
abundant means of escape in case of a sudden rise of water in 
the pitcher. But they do not always escape; Wallich records 
the presence of ‘a great number of small and harmless black 
ants, most of which find a watery grave in the turbid fluid which 
frequently half fills the cavity” +. But whatever the danger here, 
it is almost certainly greater in a nearly closed pitcher, such as 
we are considering. We can hardly imagine ants with their 
pup making their escape through an orifice 2 mm. in diameter, 
already partially filled up by roots, and through which a stream 
of water is entering. Attention is at once directed to the inner 
pitcher as a possible refuge during a temporary flooding of the 
outer one. As any of the figures will show, the opening into the 
inner pitcher is entirely on the convex side of the neck. On the 
other side the neck is concave, and provides a channel down 
which a liquid entering the orifice must perforce be directed. 
Indeed, it is difficult to see how any water can enter the inner 
pitcher until the outer one is at least three-quarters filled; and 
this holds good for any position of the organ in which water can 
* Groom, Ann. Bot. vol. vii. p. 238. 
+ Wallich, Pl. As. Rar. vol. ii. p. 36. 
