ANTS AND PLANTS 191 



velopment or growth, it is perhaps possible to steer 

 a middle course. If once we realize the ubiquity of 

 ants in the tropics, and their marauding, plundering 

 nature, we shall recognize that plants which have juicy, 

 succulent, unprotected fruits, or large fleshy organs 

 suitable for excavation such as the tubers of Myrme- 

 codia offer to ants attractions which these insects 

 will certainly not resist, and which may well prove 

 destructive to the plants. The plants, in order to 

 escape annihilation, must either develop other struc- 

 tures to protect them from ants or else must modify 

 the attractive structures to accommodate the ants, whilst 

 at the same time their original functions are pre- 

 served. It is, of course, well known that plants can 

 develop certain structures in direct response to the 

 stimuli of insect attacks : the oak-gall is merely 

 one instance out of hundreds. Such structures are 

 rightly regarded as pathological ; but it must be noted 

 that, given the same plant and the same attacking 

 insect, the structure is invariably the same. That is 

 to say, the power of producing this structure is in- 

 herited. Now, it does not seem to me a very long 

 step from this inherited power to produce a definite 

 structure in response to a given stimulus, to a power 

 to produce a definite structure in anticipation of a given 

 stimulus. That an epiphytic plant furnished with a 

 large fleshy tuber would have this tuber attacked and 

 tunnelled by ants is almost certain in lands that swarm 

 with ants, and that many of the plants so attacked 

 would perish from this extensive destruction of their 

 tissues is equally clear. The advantage of having an 

 elaborate system of galleries all ready and prepared 

 for the ants is obvious, and I see no reason why 



