ANTS AND PLANT-STRUCTURE 411 
and perhaps also in some cases to its flowers — as 
shown by Kerner — by the presence of whole armies 
of virulently - stinging ants whose very minute- 
ness renders them the more formidable. In the 
most remarkable plant-formicaria known — those of 
the Myrmecodia and Hydnophytum of the Malay 
Archipelago — the whole structure has been proved 
to be hereditary, and we may therefore conclude 
that in the Tococas of the Amazon, and other cases 
in which the cavities inhabited by the ants are 
constantly present, they are also hereditary. In 
other cases, as Spruce himself states, they are 
not so, being directly formed by the ants or being 
abnormal growths due to their irritations. 
Spruce's error was in not recognising that the 
ever-present variability in all the parts and organs 
of plants furnished the material, and the survival of 
the fittest the agency, by which these, as well as all 
other specific modifications of plants, have been 
brought about ; and that this is a far more powerful, 
as well as a more exact and certain, mode of doing 
so than the hereditary transmission of mutilations, 
the effects of which would in many cases be the 
reverse of beneficial. 
In my recent work, My Life (vol. ii. p. 64), I 
give a letter from Spruce written shortly after the 
paper was rejected, in which he explains his reasons 
for refusing to alter his paper. Three years later 
he wrote me another letter on an allied subject — 
the purport of aromatic leaves (printed at p. 65), 
at the commencement of which he says : Every 
structure, every secretion of a plant is (before all) 
beneficial to the plant itself. That is, I suppose, an 
incontrovertible axiom." 
