NEWS RELEASE

4/30/03

Mark Shwartz, News Service: (650) 723-9296, mshwartz@stanford.edu
Dawn Levy, News Service: (650) 725-1944, dawnlevy@stanford.edu

EDITORS: Greene and Gordon’s study, “Cuticular Hydrocarbons Inform Task Decisions,” appears in the May 1 edition of <I>Nature<$>. Photos and a QuickTime movie are available at http://newsphotos.stanford.edu (slug: “Ant Chat” and “Gordon”).

Relevant Web URLs:
http://ant.stanford.edu/welcome.html
http://www.stanford.edu/dept/biology/faculty/gordon.html
http://www.stanfordalumni.org/news/magazine/2002/janfeb/features/ants.html
http://www.myrmecology.org/

"Work stinks": It's more than just a slogan among ants, researchers find

Peer into the deep recesses of an ant colony and you'll discover an extremely well organized community with thousands of workers quietly going about their jobs. Some dig nests while others gather food or tend the young. Remarkably, every chore is done without supervision or direction, and some workers even switch jobs to meet the ever-changing needs of the colony.

How does an insect with a brain the size of a poppy seed decide to carry out a particular task? The answer, says a team of Stanford University biologists, has less to do with brainpower than with the ant's extraordinary sense of smell. Writing in the journal Nature, the Stanford scientists found that, when a parade of patroller ants returns to the nest, their distinctive body odor cues other workers to go out and forage for food.

This new insight into the behavior of social insects is the latest discovery to emerge from a 20-year field study of red harvester ants (Pogonomyrmex barbatus) in the southern Arizona desert -- a project designed and led by Deborah M. Gordon, an associate professor of biological sciences at Stanford.

"The question is, how does a worker know what to do?" said Gordon, coauthor of the May 1 Nature study. "There's nobody in charge, there's nobody telling it what to do."

A mature colony of red harvester ants consists of a single queen and 10,000 to 12,000 female workers. The lifespan of a worker is only one year, but a queen can live 20 years and produce thousands of new workers annually. In fact, egg-laying is the queen's only responsibility. She has nothing to do with running the colony or assigning workers to specific tasks.

Understanding the subtle cues and interactions that enable small-brained insects to build elaborate communities has become a major area of research, not only for biologists but also for engineers trying to solve intricate problems in computer science, network communications and even robotics.

"Ant colonies offer an example of a system in which the component parts -- ants -- are fairly simple and there is no hierarchical control, yet somehow the whole colony performs complex, integrated behavior," Gordon observed.

Making sense of scents

Like most ant species, red harvesters communicate by touch and smell ­ but instead of sniffing the air, ants use their sensitive antennae to detect chemical odors in their surroundings.

"Ants use chemicals the way we human beings use sight and sound. That's how they get the majority of the information about what's going on in their environment," said Stanford postdoctoral fellow Michael J. Greene, lead author of the Nature study. "Tactile, or touching, information is important to them, too, but for most ants, vision is not."

An ant's antennae are extraordinarily fine-tuned to differentiate subtle smells produced by hydrocarbons -- a naturally occurring family of chemicals that includes such well-known compounds as methane and propane. The waxy outer surface of an ant's body contains about 25 different hydrocarbons, which emit slightly different odors that are imperceptible to people, but to an ant provide important information about life in the colony.

"Hydrocarbons are simply molecules of hydrogen and carbon -- there's nothing fancy about them," Greene noted. "Yet subtle changes in the concentration of these relatively simple chemicals can produce very important and profound behavioral changes in ants."

Added Gordon: "Everybody in an ant colony shares the colony's odor, and that's how they tell one colony from another. But in harvester ants, we found a further subdivision: Different tasks smell different."

Foragers and patrollers

In their study, Greene and Gordon focused on two groups of workers with closely related tasks: "foragers," who collect edible seeds for the colony, and "patrollers," who scout potential foraging areas each morning.

"The return of the patrollers tells the foragers that it's safe to go out that day," Gordon noted. In previous experiments, she and her co-workers discovered that a single encounter with a patroller is not enough to persuade a forager to leave the nest and search for food. Rather, it takes a series of encounters with several patrollers to stimulate foraging.

"An ant has some simple rules, such as, 'I'm a forager, and if I meet a certain number or frequency of patrollers, then I'll go out,' " Gordon explained.

Foragers also interact with another group of laborers called "nest maintenance workers," who carry soil in and out of the entrance as they dig new nest chambers. But experiments have shown that encountering nest workers has no effect on a forager's decision to search for seeds.

"We wondered why it is that foragers respond to patrollers and not to nest maintenance workers," Gordon recalled. "That's when we decided to find out if different hydrocarbons are associated with different tasks."

Body chemistry

In 2001, Gordon and her colleagues analyzed the chemical content of all three worker-types and discovered that one group of hydrocarbons -- known as n-alkanes -- does, in fact, occur in different concentrations on different workers: Patrollers and foragers, who toil in the hot desert sun, have a slightly higher proportion of n-alkanes on their bodies than nest workers, who spend more time indoors.

Could this slight difference in body chemistry be the cue that distinguishes patrollers from nest workers? To find out, Greene and Gordon launched a two-year experiment designed to mimic forager-patroller interactions at nine mature ant colonies in Arizona. The scientists wanted to see if they could trick foragers into leaving a colony by enticing them with tiny beads that smelled just like patrollers.

To accomplish that, researchers would arrive at the nests early each morning and round up all of the patrollers before they had a chance to return home from their dawn patrols. "In previous studies, we found that if you take away the patrollers, the foragers don't go out at all. They just stay inside at the nest entrance," Gordon observed.

"You only have to collect about 30 patrollers to inhibit foraging in a nest," Greene added.

Thirty minutes after the roundup, researchers began dropping patroller-scented glass beads into the nest entrance. Each bead was just three millimeters wide and coated with a chemical extract that contained the same proportion of hydrocarbons as in one patroller ant.

"We used tweezers and dropped the beads in fairly slowly -- one every 10 seconds or so," Greene explained: "The idea was to drop in as many beads into the nest as patrollers that were removed, so we were really just mimicking what would happen if those patrollers were walking back into the nest."

The results were dramatic. "The foragers inside touched the beads with their antennae, and those interactions were enough to get them to come out and start foraging for the rest of the day," Greene said.

As a control, the researchers also dropped in beads coated with nest workers' extract as well as blank beads, but in neither case did the foragers come out.

Patterns of contact

"Our experiment demonstrates that a forager's response is very specific to patrollers and is not just a general response to the hydrocarbons of any ant coming back into the nest," Gordon added. "What this confirms is that an ant can assess the tasks of another ant using hydrocarbons that are specific to that task. It's not an intellectual achievement; it's a perceptual achievement. The ant doesn't have to think to get the difference between one hydrocarbon and another. It just has to have the right receptors to smell the difference."

In their study, Greene and Gordon noted that the foragers' behavior was "not a simple response to patroller extract alone." To trigger foraging, the scented beads also had to be dropped into the nest at a pace that the foragers seem to recognize.

"An ant responds to the rate at which it meets ants of another task," Gordon explained. "Meeting one patroller every 10 seconds seems to be sufficient to elicit a response, but in real life, the experience of each ant is very variable. Our next step is to figure out what difference the rate of contact makes."

Ants do not tell each other what to do when they meet, she emphasized: "What seems to matter to an ant is the pattern of interactions it experiences rather than a particular message or signal transferred at each interaction."

A growing number of high-tech engineers are also interested in deciphering those patterns and applying them to such fields as telecommunications, computer networking, artificial intelligence and robotics. In fact, earlier studies on ant interactions by Gordon and other scientists have inspired engineers to design robots capable of interacting and performing tasks in unpredictable situations. Research on ant colonies is even being used to unravel secrets about the inner workings of the mind.

"A brain consists of units called neurons, none of which can make any global assessments, and yet somehow in the aggregate, lots of neurons together manage to think and remember," Gordon explained. "We hope that by learning more about how a particular system -- like an ant colony or a brain -- is organized, we'll understand more about how all such complex systems work."

Last year, Gordon received a three-year interdisciplinary research grant from the university's Bio-X program to explore how colony size affects network behavior. Meanwhile, it remains to be seen whether the results of the Nature study will apply to the other 12,000 known species of ants. The Gordon lab is currently analyzing the hydrocarbons of Argentine ants ­ an imported variety that has all but wiped out red harvester ants and many other native species in the Western United States.

Argentine ants are notorious for invading homes and apartments during periods of heavy rainfall. Perhaps one day researchers will identify the hydrocarbons that tell unwanted foragers when it's time to leave.

"That's the question we get most often: How do you get rid of those Argentine ants in my kitchen?" Greene said. "If we had the solution, we'd be doing all right."

Greene's work was supported by a National Research Service Award granted by the National Institute of Deafness and Other Communicative Disorders -- one of the National Institutes of Health.

-30-

By Mark Shwartz