52 
ANIMAL INTELLIGENCE. 
consequently it cannot Lave been tbe mere sight of a larva 
which had induced them to follow her. 
Further experiments proved, as we might have ex- 
pected, that although an ant is able to communicate to 
her friends in the nest that she has found treasure some- 
where outside, she is not able to describe to them its pre- 
cise locality. Thus, having exposed larvae and placed an 
ant upon them as before, Sir John watched every time she 
came out of the nest with friends to assist her, but instead 
of allowing her to pilot the way, he took her up and 
carried her to the larvae, allowing her to return with a 
larva upon her own feet. Under these circumstances the 
friends, although evidently coming out with the intention 
of finding some treasure, w'ere never able to find it ; but 
wandered about in various directions for a while, and then 
returned to the nest. Thus, during two hours she 
brought out in her successive journeys altogether no less 
than 120 ants, of which number only 5 in their unguided 
wanderings happened to find the sought-for treasure. 
This result seems to prove, as we might have expected, 
that the communication is of the nature of some sign 
amounting to no more than a 6 follow me.’ Other experi- 
ments confirmed this result, and also brought out the 
fact that ‘some species act much more in association than 
others — Formica fusca > , for instance, much less than 
Lasius niger .’ Thus Sir John Lubbock placed some honey 
before a marked specimen of the former species ; but 
although she visited and revisited the honey during an 
entire day, she brought out no friends to share it ; and 
although in her journeys to and from the nest she happened 
to pass and repass many other individuals, they took no 
notice of each other. 
The obvious objection to these experiments, that an 
ant observing a friend bringing home food or a pupa 
might infer, without being told, that by accompanying 
the friend on the return journey she 6 might participate in 
the good things, 5 has been partly met by the fact already 
stated, viz., that there is so very marked a difference in 
the result if, on experimenting on two ants, one had access 
to a large treasure and the other only to a small one. But 
