230 Groom . — On Dischidia rcifflesiana (Wall.). 
the concave leaves of D. Colly ris (Wall.). In the literature 
there appear to be no definite statements as to whether the 
ants merely utilize the pitchers as homes, or whether they 
bring material with which to build nests. Treub, however, says 
‘ Les ascidies sont devenues de veritables nids de fourmis.’ 
The next question which requires an answer is, For what 
purpose did the pitchers originally arise? Treub has worked 
out the development of the pitchers ; the ontogeny of the 
pitchers supports the view of the evolution which is obtained 
from a comparison with other species of Dischidia^ . There 
are many species in which the leaves are concave on neither 
face ; a form of D. nummularia , which I saw growing abun- 
dantly about Singapore, possesses small fleshy leaves which 
are biconvex in section and afford no shelter to the roots. 
D. cochleata (Bl.) has leaves with a convex upper surface and 
concave, pale purple, lower surface : the leaves in this case 
tend to diminish the loss of water by the roots. In D. Collyris, 
Wall. ( Conchophyllnm , Bl.) the densely set leaves, lying close 
over the surface of the host-plant, have concave purple lower 
surfaces, and they completely screen the roots ; in the spaces 
under the leaves ants build nests and bring hither stores of 
humus (small fragments of leaves, &c.), for the purpose. 
Finally some of the leaves of D. rafflesiana have their lower 
surfaces so concave, and their margins curved into the con- 
cavity, to such an extent, that the pitcher-form is the result. 
In addition the roots in connexion with the pitchers altogether 
desert the surface of the host and ramify within the cavities of 
the pitchers. There is little reason to doubt that originally 
the concave leaves protected the underlying roots, as they do 
in D . Collyris : and as do the flattened pseudo-bulbs of the 
epiphytic Oncidinm Limminghii 2 and the flattened stem of 
the epiphytic Poly podium schomburgkianum 2 . 
But with regard to the final stage in the evolution of the 
pitchers, five views may be propounded. It may be suggested 
that, having attained the concave form now seen in D. Collyris , 
1 See Beccari, loc. cit. 
3 Goebel, Pflanzenbiologische Schilderungen, i, pp. 228-229. 
