IN JAVA. 



81 



can effect it) ; thtit the jKtUen graii^s exsert tlieir pollen tu}je^ 

 while still in the anthers; and that both the external and the 

 internal surfaces of the lobes of the pistil are covered with 

 pupilliB, indicating that these surfaces are functionally active. 



I have never observed these flowers approached by the ants 

 that infest the interior, nor by any oth^ insect, which to gain 

 admission to the flower, even if open, must be very small 

 indeed. The anthers and the pistil do not seem to reach 

 maturity together, yet it would seem that self- fertilisation 

 alone can take place ; perhaps the tnbes of the poUen grains 

 which fall to the bottom of the corolla manage to reach the 

 lower lobes of the pistil and produce fecundation. 



The seeds I planted germinated with great freedom, and I 

 cultivated quite a number of young Mijrmemdia^ whose growth I 

 watched with the greatest interest. 3Iany of them I kept qnito 

 isolated from the interference not only of the Fheidde Javana, 

 which seems to be the only species of ant wliich lives in these 

 plants in their native state, but of all other species, and I was 

 surprised to find that from their very earliest appearance thh 

 curious (jdlleried strmture arose mthout the ^eseme of the 

 anfs, and that the plants continued to grow and thrive vigo- 

 rously in their absence as long as I cultivated them. Some 

 bulbs had a single canal reaching to their centre from a round 

 orifice opening generally close to the little tap-root; others 

 presented one or two locnJi in the interior, without any 

 communication at first with the exterior, 

 partially full of a spongy substance look- 

 ing like its owTi degenerated tissue. These 

 chambers invariably developed a spongy 

 pith — which in a section it was not dilli- 

 cult to trace out in advance in the still 

 fleshy substance— towards and to open at 

 last at one or more spot, on the exterior of "™4o™cii'«S? 

 the bulb. Secondary galleries, arising in olbeh one, 

 the same manner as the primary, soon formed communicating 

 channels, extending with age, throughout the whole of the 

 growing bulb. At a later period, in Amboina, where the 

 Myrmecodia and the H^dmphytum were very abundant, I 

 found many specimens containing a large central and quite 

 isolated chamber full of water— not niin- water— round which 



