54 
Our myrmecophilous plants are, without exception, epiphytes. As 
such, they are exposed to dearth of water and dearth of mineral food. 
When they protect themselves against injury by the former by using 
devices to reduce the- transpiration, they aggravate the latter difficulty. 
Epiphytes have many ways of overcoming their difficulty of obtaining 
mineral food, such as the maintenance of remote ground-connections; 
parasitism; complete exhaustion of their own dead parts; cooperation in 
the accumulation of an aerial “soil/ 7 in the mossy forest and in tree-top 
gardens at lesser altitudes; special humus-collecting structures, such as 
have just been described; the insectivorous habit, in Nepenthes, and 
the attraction of insects for the sake of the debris they bring, or for 
their excreta or their carcasses, as is the case with the plants now under 
discussion. The plants waste none of their parts to support the ants, 
offering them only a tolerably moist shelter, and this is very evidently 
a sufficient inducement for the ants to seek them, for I have never 
found a healthy individual of one of these plants without its tenants. 
The latter are not specialized in adaptation to their specific hosts, for the 
same ants inhabit the chambers of different plants ; for instance I have 
found one kind in Polypodium sinuosum, Myrmecodia , and Hydnophy- 
tum all in a single tree. Although ants have not the reputation of being 
untidy housekeepers, the chambers which they occupy are never really 
clean. The plant can of itself effect the quick removal of liquid ejecta; 
if can get rid of solid ones, only as they are dissolved. I have found a 
fungus in an apparently healthy Polypodium sinuosum , growing in the 
lining of the chamber- and at first imagined that it might be analogous in 
function to mycorhiza, but it is not always present and it was probably 
merely accidental. Both of these ferns are without other roots than 
such as are necessary for their firm attachment and they habitually 
grow on bare branches, without any mass of other epiphytes; therefore, 
they would be in especial straits for mineral food if it were not for their 
tenant ants. Nevertheless they are conspicuous for the very ready falling 
off of their leaves, conclusive evidence that they are not in practice obliged 
to husband their ash-constituents. The facts that Polypodium sinuosum 
can live after its chamber is plugged (Goebel), and that Hydnopliytum 
and Myrmecoidea can grow and develop their chambers without the pres- 
ence of ants (Treub), do not prove that the ants are useless to the plants 
any more than the power of Drosera to live under favorable conditions 
without insects is a demonstration that the plant is not insectivorous. Of 
the two ferns, Lecanopteris is the more highly developed in myrmecophily, 
not only in grosser, conspicuous characters, but also in the perfection of 
its chamber, the walls of which, as described and figured by Yapp, are 
made up of pockets, which are doubly serviceable as collectors of possible 
food, and as increasing the absorbing area. 
The doctrine that these stems are enlarged as water-reservoirs, and 
chambered and the reservoir-tissue removed because they are too fleshy, 
