Ant Encounters: Interaction Networks and Colony Behavior

In The Adventures of Pinocchio, Master Antonio taught ants their ABCs. It was the right thing to do; after all, ants are already proficient in arithmetic. In my experiments with Boris Ryabko, red wood ants have demonstrated their ability to add and subtract small numbers, and in Deborah Gordon's study, ants have been shown to be able to estimate quite precisely the number of encounters they have with their nestmates.

Gordon has spent many years in the Arizona desert studying harvester ants. These assiduous creatures collect seeds and store them in their nests as food for the colony. Perhaps it was they who helped Cinderella to collect and sort out seeds: in the real world, ants amaze us even more.

In her new book, written in a lively style and accessible to a general audience, Gordon describes the sophisticated experiments that led her to intriguing insights about how an ant colony can solve vital problems, such as finding the shortest path to the best food source, allocating workers to different tasks or defending a territory from rivals. She observed hundreds of tracks made by individually marked ants; she enticed ants to switch their tasks by dropping glass beads, some coated with patroller scent and some with maintenance worker scent; she even discovered which gene has to be activated or deactivated to enable ants to go outside a nest to forage.

Ants use the rate of interaction to make task decisions, Gordon says, and so they "reason" something like this: "If I meet a returning forager every 5 seconds, I am likely to go out again within 15 seconds. But if the returning foragers show up every 2 seconds, I am likely to go out again within 10 seconds."

Ant colonies have different levels of collective sensitivity to encounters, and this dramatically influences their prosperity. Natural selection shapes the regulation of foraging, and differences among colonies lead to differences in reproductive success.

Each ant must contribute to the wealth of its "alma mater". To what extent can an individual ant be said to be aware? Gordon contends that ants are not smart but ant colonies are, and the basic mystery about ant colonies is that there is no management. In line with many other researchers studying social insects and other creatures that react collectively to challenges from their environment, Gordon adheres to the "swarm intelligence" concept, by which patterns of interactions among individuals are governed by rules of self-organisation, and fairly simple units generate complicated behaviour by the group as a whole.

One cannot help but concur with Gordon that no single ant grasps the big picture even as each one contributes to the group's success. But although the author compares the modes of life in different species, including tropical weaver ants, forest red wood ants and so on, she focuses chiefly on her favourite, harvester ants. Indeed, members of this species are representative of the behavioural "mainstream" among the approximately 12,000 ant species on Earth. In their colonies, individuals essentially act as replacement parts.

But in some species, a completely different kind of social organisation has been discovered, based on cognitive specialisation. For example, in red wood ants, unlike foragers, only scouts can do clever things such as remembering sequences of turns towards the goal, noting irregularities and even doing simple arithmetic. What is even more interesting, this is their lifelong role. Knowledge itself is power but ... not for all.

So when you accidentally crush an ant, it may mean nothing for the interaction network of which it was part, but it is possible that you have deprived a community of a member of an elite club of rare "cognitive specialists".