Networking ants: how do ants find the cake crumb you dropped?

As everyone knows, where there is a picnic, there will be ants. How many show up to share your sandwich depends on how quickly one discovers the first morsel and alerts its nestmates to come share in the loot. But how do ants find food in the first place? Do they stake out all the prime spots before you get there?

Ants cannot see well enough to distinguish one object from another; they find food only when they come close enough to smell it. This presents a problem for the ant colony as a whole: How should each ant move around to maximize the probability that if food appears somewhere, some ants will be there to recognize the scent?

To visualize the solution to this problem, imagine the paths of the patrolling ants as a network of lines. If each ant were to walk in a straight line, the net formed by all the paths would resemble a bunch of straight sticks thrown on the ground. If each ant were to walk in a very wiggly, convoluted line, the net would resemble an assortment of shoelaces or pieces of string. Which searching method is more effective for a foraging ant colony?

With the help of Fred Adler, a mathematical biologist at the University of Utah, I set out to find which path shape would lead the colony to discover the most bits of food. Fred devised a mathematical model and computer simulations of possible patrolling paths. According to our model, the best searching tactics vary with the density of searchers. When searchers are few, relatively straight paths are better than wiggly ones, which may go round and round in the same place and leave large areas unexplored. But when the entire area is crowded with searchers, wiggly paths are fine. Each individual makes many turns within a limited radius. But if food appears outside an ant's narrow purview, another searcher will probably find it.

I decided to compare our mathematical model of search patterns with the actual behavior of Argentine ants. (This South American species, which has now invaded many of the world's warm coastal areas, frequently demonstrates its brilliant food-finding talents in California, including in my own kitchen.) I gave captive ant colonies access to enclosed arenas with just enough food aroma (a layer of dried sugar water) to draw them out of their nests, but no morsels to carry back. The ants responded by searching the arena and then going back to their nest boxes. I did this experiment many times, varying density by changing both colony and arena size and videotaping the ants using a computerized image analysis system developed by programmer Garr Updegraff. I then traced the paths of an the ants (see examples below). I found that they adjusted their path type according to the number of searchers, adopting the shapes that Adler's model had predicted. When there were few ants in a large space, the ants moved in straighter lines. When the ants were more crowded, their paths became more wiggly.

Without any central coordination, each ant somehow adjusts its path to the number of searchers. The overall patrolling pattern resembles an elastic net, whose lines stretch and straighten when the ants are few but contract into convoluted curls when they are plentiful. Result: Under either set of conditions, the whole group's chances of finding food are maximized.

COPYRIGHT 1997 Natural History Magazine, Inc.
COPYRIGHT 2008 Gale, Cengage Learning