XXIV ANTS AND PLANT-STRUCTURE 393 
ants, for it is a very common thing all through the 
order of Melastomes. In species which have the 
leaves of each pair nearly equal, it is usual to see 
some of the smaller ones saccate and others alto- 
gether esaccate on the same plant. [I have often 
examined half-grown plants a7td have seen that sacs 
begin to be developed {by inheritance) long before 
ajiy ants touch thein, but that when the sacs are 
taken possession of by ants they speedily became 
much enlarged.] 
Seeing, then, how the sacs on the leaves have 
originated, and what purpose they serve, it is plain 
that a species of Tococa, like T. planifolia, inhabit- 
ing the very river's brink, and liable to be com- 
pletely submerged for several months of every year, 
could never serve as a permanent residence for ants, 
nor consequently have any character impressed on 
it by their merely temporary sojourn ; even if their 
instinct did not teach them to avoid it altogether, 
as they actually seem to do ; whereas the species of 
Tococa growing far enough inland to maintain their 
heads above water even at the height of flood are 
thereby fitted to be permanently inhabited, and are 
consequently never destitztte of saccate leaves, nor at 
any season of the year clear of ants ; as I have 
reason to know from the many desperate struggles 
I have had with those pugnacious little creatures 
when breaking up their homes for the sake of 
specimens. 
In one species (hb. 3477) with seven -ribbed 
leaves, growing by the Rio Negro near the mouth 
of the Casiquiari, the leaves on some plants have a 
small distorted sac at the base inhabited by ants, 
and on others are nearly all esaccate ; and I noted 
