23 6 Groom. — On Disc India rafflesiana ( Wall .). 
roots of two sorts. Schimper *, in his masterly work on epi- 
phytes, points out that there are two classes of epiphytes which 
possess roots of two kinds. First there are those which have 
negatively heliotropic roots, which are simply organs of 
attachment : and other roots which are organs of absorption, 
the latter being positively geotropic and descending to the 
ground. Secondly there are epiphytes which have negatively 
heliotropic roots which, though primarily organs of attach- 
ment, are to a certain extent absorptive : and, in addition, 
other roots which are negatively geotropic and purely 
absorptive. D. rafflesiana forms a third class of heterorhizal 
epiphytes ; for it possesses attaching-roots which are also 
absorptive, as well as the pitcher-roots which are purely 
absorptive but possess no marked geotropic properties 1 2 . 
Neither is it a unique phenomenon that the absorptive roots 
should be associated with peculiar humus-collecting leaves, 
for the same is the case with heterophyllous epiphytic Platy- 
ceriums. The analogy between these Ferns and D. rafflesiana 
is rendered still more striking from the fact, discovered by 
Goebel 3 , that ants nest amongst the humus collected in the 
mantle-like leaves of Platycerium alcicorne . 
As the external conditions of the two sets of roots of 
D. rafflesiana are dissimilar — one set creeping over the shoots 
of the host and the other set hidden from direct sunlight — ■ 
and, as the pitcher-roots have no attaching function, we should 
expect to find corresponding differences in structure. And 
our anticipations become all the stronger by reason of the 
fact that other epiphytes display anatomical distinctions in 
their two sets of roots 4 . The pitcher-roots do differ in 
structure from the attaching-roots, but the differences are not 
1 Schimper, loc. cit. 
2 The pitcher-roots appear to be negatively heliotropic, but no experiments 
have been made showing that this is the case. 
3 Goebel, loc. cit. It is not known whether the ants simply utilize the particles 
of vegetable matter, &c„ contained in the niches formed by the mantle-leaves, or 
whether they take thither stores of vegetable fragments, earth, &c. In the latter 
case the resemblance to the state of affairs in D. rafflesiana would be very close. 
4 See Schimper, loc. cit. 
