t8o NATURAL SCIENCE. Sept., 



Dischidia rafflesiana. Some of its leaves are hollowed out into pitchers. 

 It has roots of two sorts — attaching and nutritive. The latter are 

 given off into the cavities of the pitchers, within which they ramify. 

 Treub (7) was of opinion that these pitchers functioned solely, or 

 mainly, as receptacles for rain-water and economisers of the water 

 transpired within the pitchers. He showed that the epidermis lining 

 the pitcher-cavity could not absorb water. But I have shown that 

 very considerable amounts of humus-providing materials are collected 

 in these pitchers (8), and that the roots within the pitchers absorb 

 this food placed at their disposal. These pitchers have the same 

 function as the rosettes of Anthurium or the "pocket-leaves" of ferns 

 — they collect humus. However, they are able to collect more water 

 than is possible to their simpler analogues. One curious fact about 

 these pitchers is the constant presence of ants living within them. 

 Probably a large portion of the food found in these pitchers is con- 

 veyed thither by the ants. The plant provides shelter for the 

 ants, and they in return bring these materials (for the purpose of 

 building nests). 



Externally unlike Dischidia rafflesiana, but very similar in funda- 

 mental design, are two Rubiaceous Epiphytes — -Myrmecodia and Hydno- 

 phytum. In each of these genera the base of the stem takes the form 

 of a great swollen tuber, which is hollowed out by a series of galleries 

 or canals continuous with one another, and communicating with the 

 exterior by, mostly small, apertures. These galleries are always 

 found to be occupied by innumerable ants belonging to a certain 

 definite genus and species. In a typical case, lenticels occur in the 

 walls of the galleries from which there protrude, more or less, distinct 

 roots. In principle, then, this great tuber corresponds to a pitcher 

 of Dischidia ; the stomata of the epidermal lining of the pitcher corre- 

 spond to the lenticels opening into the galleries; and the roots 

 protrude into the galleries and the pitcher-cavity respectively. Rain- 

 water can only pour into some of the pitchers of Dischidia, because 

 their mouths sometimes point downwards ; but rain-water can enter 

 freely into the galleries in only one known species. In most cases, the 

 external openings of the galleries are so small that water could only 

 enter in the form of a thin capillary film. Formerly it was supposed 

 that the ants constructed the galleries, and that these animals were 

 essential to the existence of these plants ; but Treub (9) showed 

 that the galleries developed in the absence of ants, and that the 

 plants could exist without the proper ants. The hypothesis that the 

 galleries are respiratory passages is quite inadequate. The whole 

 arrangement has been looked upon as a mechanism for the economy 

 of water. This hypothesis is backed up by no experiments, and, at 

 present, is only bolstered up by a few facts, and more or less distant 

 analogies. 



It is still possible that the ants have been the prime factors in 

 the evolution of these excavated tubers. The development of the 



