Skip to content


Life in the Colonies

Ants carve underground labyrinths, gather food, nurture their young and adapt to all sorts of environmental changes—with no leadership whatsoever. What drives this 'swarm intelligence'? A Stanford biologist thinks she's close to some answers.

Edward McCain

RAINBOW COALITION: Tiny dabs of paint identify individual ants.

View photo album >>

By Mitchell Leslie

Crouched beside a scrawny bush, Deborah Gordon jiggles a grass stem inside a dime-sized opening in the ground. “Would you like to be colony 961?” she asks with a laugh. On this bright, baking August morning, Gordon is hunting for one of the nests of red harvester ants that she and her helpers have been tracking on a Stanford-owned patch of desert in southwestern New Mexico. Across the valley to the west, the cliffs of the Chiricahua Mountains blush in the early sunshine. The air is silent except for the piping of small birds and the whoosh of an occasional car on the highway, half a mile away.

After a little probing, a bemused ant pops from the hole. It’s not a red harvester; another species lives here. Gordon, an associate professor of biological sciences, consults a map showing the locations of all 300 red harvester colonies currently nesting on the 25-acre research site, then briskly walks away, still searching for the elusive nest among ankle-high grass and low shrubs.

The lost colony, which eventually turns up nearby, is small and young. But some nests that Gordon, MA ’77, and her helpers are checking out today are old friends that were here during her first visit in 1981. For 20 summers, she and a changing crew of undergrads, graduate students and postdoctoral fellows have tramped through this plot about 120 miles southeast of Tucson, Ariz., keeping tabs on the harvester nests. They have precisely mapped the coordinates of each nest, recording the founding of new colonies and the demise of old ones. To understand how red harvesters respond to the vagaries of their environment, the researchers have bothered nests by stacking toothpicks on the mound, blocked foraging trails and spirited away unsuspecting ants with tiny vacuums. Sometimes Gordon’s studies call for high-tech methods like DNA analysis and fiber-optic microscopes. Sometimes they call for a backhoe.

Her goal sounds fairly simple. She wants to know how an ant nest functions, how a bunch of dumb insects can achieve such complex, seemingly coordinated behavior, and how that behavior might change over time. What makes her work challenging is that ant societies, unlike most human organizations, have no leaders, no chain of command. As Gordon puts it, “the basic mystery about ant colonies is that there is no management.” Ants work together to sculpt exquisite underground chambers, care for their young, store food for hard times and cope with environmental change—all without any apparent direction.

Nobody knows exactly how it happens. But according to Gordon, we may soon understand how at least some of this intricate, communal behavior emerges from interactions between individual red harvester ants. “We are close to knowing enough to explain the behavior that we see,” she asserts.

Understanding ants is important in its own right. Though most of us ignore them unless a raiding party storms our picnic, these ubiquitous creatures, whose species number around 15,000, have an enormous impact on the planet’s ecology. One scientist has estimated that the total weight of all the ants on earth compares with the total weight of people. Ants of all kinds gobble up vast quantities of insects, churn more soil than earthworms and spread the seeds of many species of plants.

Ant societies might also provide clues to the workings of other complex natural systems. Cells in a developing embryo, plants and animals in a rainforest, and neurons in the brain share a common trait: they function as a biological unit without discernible direction. And while an ant nest differs in important ways from a brain or an embryo, what we learn about its organization might hint at how a thought is born or how a ball of nondescript cells morphs into a baby. Moreover, Gordon’s findings about the flexibility of harvester ant society are inspiring researchers in other fields to design more adaptable robots, smarter computers and a more efficient Internet.

Try to imagine a winning football team without a coach, a polished platoon without a sergeant, a thriving country without a government. The notion of a complex, leaderless and successful organization defies our experience. However, as Harvard evolutionary biologist and ant expert E.O. Wilson points out, ants discovered communism more than 100 million years before Marx—and they made it work. While movies like Antz would have us view ant colonies as societies dominated by a ruthless soldier caste, the leaderless structure was apparent even in biblical times. In Proverbs, for instance, Solomon admonishes, “Look to the ant, thou sluggard . . . Which having no chief, overseer, or ruler, Provides her meat in the summer, And gathers her food in the harvest.”

What about the ant royalty that the rest of the harvesters fuss over? First off, there is no king. As for the queen, she is a plump egg-laying machine with a life span of about 20 years who extrudes as many as 10,000 eggs a year. Her royal chamber, deep within the nest, is a delivery room rather than a command post.

Unlike some human royals, the ant queen does perform an essential function. She gives birth to the workers—an all-female force that builds, maintains and guards the nest, tends the young and gleans small seeds from the dusty soil. Because workers are sterile and live only one year, and the nest has only one reproducing queen, her death means the colony will die out a year later. But when you consider that the workers go on as usual during that final, queenless year, it’s clear she didn’t rule.

Depending on the queen’s maturity (reproductive volume ramps up to full capacity at age 5), a red harvester nest might hold a few thousand to 10,000 or more adults, with the queen and workers making up about 98 percent of the colony. The remaining nestmates—a fleeting minority—are males and “virgin queens.” These fertile individuals fly out of the nest for a single day to mate with peers from other colonies. The males are nothing more than sperm-delivering missiles: they cannot feed themselves, and they die right after mating. The virgin queens are winged daughters of the queen. After mating, they shed their wings and settle down to start a colony of their own, storing the sperm they’ve just collected to produce decades’ worth of progeny.

The day-to-day labor of the nest thus falls to the workers. Without any oversight, they achieve marvels through teamwork. You can see this best by looking at their home. Red harvesters don’t waste much effort building above ground. Their “hill” is a broad mound an inch or two high, strewn with fine gravel. A large, old nest might be three or four feet across, with foraging trails as wide as bicycle tracks wending 100 feet into the surrounding vegetation. It’s hard to appreciate the nest’s underground structure without digging it up. Gordon occasionally finds such excavations necessary—there’s no other way to determine a colony’s population and get the nest back to her campus laboratory for continued study—but she hates to demolish the ants’ hard work.

Gordon has observed that harvester ant workers, though less than half an inch long, scrape out an intricate network of tunnels and chambers extending more than six feet down through the concrete-hard soil. The architecture resembles a child’s ant farm (minus the plastic silo and barn) but on a much larger scale. Possibly to protect the colony from floods, the diggers may drill an escape tunnel that runs several feet farther into the rocky layers below. Helping regulate the internal climate of the nest, ants line each chamber with clay that hardens like adobe. In fact, Gordon says, these chambers look remarkably like some of the rooms the Pueblo Indians carved into the rock at the nearby Gila Cliff Dwellings National Monument.

Impressive achievements. But zoom in on toiling ants and you get a different picture, she says. Up close, individual ants often seem shiftless or downright incompetent. Solomon might retract his biblical advice if he could peer inside a harvester nest with a fiber-optic microscope, as Gordon has done. What she sees—apart from some frighteningly magnified insects dashing past the lens—are plenty of ants just standing around. These “idlers,” as Gordon calls them, are not necessarily lazy; they could be waiting to work, or they might constitute a reserve force for rare emergencies. However, their presence belies the ant’s reputation as a model of diligence.

And even busy ants aren’t necessarily candidates for Insect of the Month. For example, Gordon and her colleagues collected red harvesters returning from food-gathering trips and encouraged them (by tapping their heads with a small stick) to release their prize. A surprising number of these ants, they found, carried useless husks and other junk.

“I see little that seems efficient about the ways that ants forage or interact with their neighbors,” Gordon writes in her book, Ants at Work: How an Insect Society Is Organized (Free Press, 1999). “I am never struck by their perfection.”

Which leads us back to the fundamental question: how does what Gordon calls “a collection of inept individuals” manage to keep the colony running so well?

Answering that question wasn’t her original aim. A French major at Oberlin College, Gordon didn’t get hooked on biology until her senior year, when she took a class in comparative anatomy. “It was my first glimpse of the idea of natural order,” she recalls. “I thought you found order in things that people created, like symphonies and fugues.” After fortifying her credentials with a Stanford master’s in biological sciences, she went to Duke to work in animal behavior but found herself even more intrigued by the phenomenon of embryonic development. That’s when ants captured her interest.

It wasn’t the intellectual leap it might seem, Gordon says. Just as no ant rules the mound, no master cell choreographs an embryo. Instead, the changes that put the eyes, heart, kidneys and other organs in the right place depend on interactions between cells. Today, Gordon says she’s probably happier studying ants instead of cells because “I like to see what I’m studying.” What’s more, she says, ants reveal two levels of behavior simultaneously: individual and colony. “Zoom in, you see the ants; zoom out, you see a colony,” she writes in Ants at Work. “Ants and colonies are both there in front of you, all the time.”

Her first field studies bombed—almost literally. She picked a study site close to Duke: a nature reserve adjacent to the military base Fort Bragg, where sandy soil allowed two people with shovels to dig up an ant nest in five minutes. (In New Mexico, excavating a nest is a grueling project that begins at dawn with a backhoe clawing a trench alongside the colony. The ordeal ends hours later, with the frustrated, overheated, grimy scientists scrabbling to find the queen’s escape tunnel in the bottom of the six-foot pit.) Gordon found it disconcerting, however, when artillery shells from the fort’s firing range landed not too far away, prompting worries about the gunners’ accuracy. And then one day the downdraft from a landing helicopter blew away the paper strips she had meticulously set out for an experiment. The next year, she came west.

In the desert, the ants come out to work shortly after sunrise and seek shelter when the ground temperature climbs above a foot-frying 125 degrees. Often, that means the ant-watchers must put in a full day’s work before noon. Just after 5 a.m., under a violet pre-dawn sky in early August, Gordon and her five helpers rendezvous in the school-style cafeteria of the Southwestern Research Station. Located in a beautiful canyon in the Chiricahua Mountains, this former guest ranch is part motel, part science lab, part summer camp and has been Gordon’s base since the beginning. The station was once the Western playground of the Rockefeller family, but now it’s run by the American Museum of Natural History, in New York—which explains the wire-mesh Statue of Liberty that overlooks the pool.

After a self-serve breakfast of fruit and cereal, we hop into a creaking white Suburban and drive down the pocked, twisting road into the desert. Everyone here wears basically the same uniform: wide-brimmed hat, long-sleeved shirt, long pants tucked into high socks that reach above sturdy hiking boots. The outfit protects against the roasting sun and keeps the biting, stinging ants out of their pants—most of the time. Despite all precautions, the researchers do occasionally get zapped. “Usually it happens when I’m standing next to a nest telling my students not to be stoic,” says Gordon. “It really does hurt.”

Before 7 a.m., nests all over the site are seething with square-headed ants the color of crisp bacon. As the colonies get busy, so do the researchers. Gordon’s two undergraduate helpers, senior Jennifer Chu and junior David Garfield, sit on folding camp chairs, intently watching the entrance of a large nest. As ants flow in and out, Chu calls out some kind of code.

“Out B B S with something.”

“In O O O.”

The letters stand for the colored dots the team has applied to 150 ants. A few days ago, they chilled the insects in an ice cream maker and applied a tiny dab of paint to the head, thorax and abdomen of each before returning the labeled ants to their nests. Today they are recording the ants’ activities, recognizing individuals by their combinations of blue, orange, silver, green and other colors.

Sitting in the sun for the next four or five hours, Chu and Garfield will take turns watching the ants and writing down the results. “We’ve done it up to five hours—but after five, your brain really goes to mush,” Chu says. They are trying to determine whether ants that have ventured out to forage tend to go back out to forage again and again, or instead tarry in the nest to assess how much the colony has amassed before leaving to gather more.

Postdoc Mike Greene, a recent arrival from Oregon State, is out there, too, watching how ants respond to different scents. He sets a handful of tiny glass beads outside the entrance to a nest. Some beads have been coated with the colony’s own scent, others with the scent of another harvester colony. The foreign smell enrages the emerging ants. They bite the bead and attempt to haul it away, or else climb aboard and try to sting it.

Zia Khan, a computer science undergrad from Carnegie-Mellon University, is filming ants as they walk along trails, using a digital camera mounted on a crossbar between two tripods. Khan’s supervisor is developing a program to “watch” film and identify which tasks ants are performing, thus sparing researchers many hours of monotonous work.

During the morning, everyone takes a turn searching for colonies and marking them with rocks painted with green numbers. By 11:30 a.m., the ants have retreated to the coolness of their nest, and the crew retreats to the research station for lunch.

This has been a typical day in the field—a routine followed by Gordon and her group six days a week for six weeks a year. In addition to her crew of students and helpers, she brings her family along on these trips: husband Ben Crow, a sociology professor at UC-Santa Cruz, and their children, Sam, 6, and Eleanor, 3. The kids usually remain at the station with Ben during the day. Though Sam has visited the research site, curious little Eleanor definitely has to stay away. “Eleanor,” says Gordon, “is the type of person who would pick up an ant and get bitten.”

After two decades of stalking red harvester ants in the desert, eyeing them in the lab and building mathematical models of their activities, Gordon believes she has pieced together a fairly complete picture of how they work outside the nest. One key to their success, she finds, is their flexibility.

Unlike some ant species, red harvester workers lack physically distinct groups that perform a single task throughout life, such as defense or food gathering. Instead, a harvester can have many jobs during her life. This “task switching” seems to occur in a loosely set sequence. The ant usually starts out laboring within the nest, maybe tending the young, and later moves to outdoor tasks, such as working on the colony’s midden, or garbage dump. The last task she will perform in her yearlong life cycle is foraging.

Gordon discovered, to her surprise, that the colony as a whole also changes its behavior over time: in human terms, it acts more grown-up. She observed that when foragers from two adolescent colonies (nests aged 2 to 4 years) meet, they often fight and will return to the same spot day after day, tussling every time. In contrast, when foragers from older colonies (5 to 20 years or more) run into one another, they will usually stay away from the brawl site in the future.

Such flexibility could offer important advantages. In a young, rapidly growing colony, the production of hungry grubs outpaces the workforce that will feed and care for them. The 4,000 workers of, say, a 3-year-old nest must gather enough food to prepare for 6,000 young, whereas the 10,000 larvae in a 6-year-old nest have 10,000 workers to feed them. Thus, the inhabitants of the mushrooming adolescent colony might need to be more aggressive to round up an adequate supply of seeds.

But how would they “know” which behavior—in this case, aggression vs. avoidance—is right for their colony? With workers surviving only a year, it’s not as though the ants are benefiting from the wise counsel of grannies with gray antennas.

Like other scientists who study the social insects (ants, bees, wasps, termites), Gordon makes no claim that the individual insects have the brainpower for such complex assessments. Rather, she thinks ant behavior is governed by simple rules—rules that would describe how an ant’s actions spring reflexively from interactions with other ants and from constant evaluation of the physical environment. For example, Gordon and her helpers found that foragers will not come out to hunt for food unless they first meet patrollers, the ants that emerge first in the morning and reconnoiter. Remove the patrollers early enough, and the colony stays in bed all day.

Through observations such as these, Gordon believes she has identified a link between individual and colony behavior. Each ant’s “decision” on whether to become active and what activity to perform seems to depend not on a single meeting but on the patterns of interactions among many individuals, she says. Specifically, red harvester ants seem to be recognizing and “tallying”—not numerically, but through sensory cues such as scent—the other ants they encounter and adjusting their behavior accordingly. If the foragers meet a certain number of patrollers within a certain time, for instance, they may “decide” to begin foraging. And preliminary work from Gordon’s lab suggests that the more midden laborers an ant meets on a given morning, the greater the likelihood that she will labor on the midden herself that day.

Gordon thinks she has now compiled enough data to step back and view the harvester ant colony as a complex system. “The main problem remaining,” she asserts, “is how to put together what we know about individual behavior and figure out how the whole system ticks dynamically.” Her next step, she says, is to build computer models mimicking the moment-to-moment interactions between ants to see if she can simulate the behavior of real nests.

That may be premature, argues Thomas Seeley, a behavioral ecologist at Cornell University. While Gordon’s emphasis on the flexibility of ant and colony behavior has been helpful, he says, most of her ideas have yet to be established. And no one knows enough about what goes on inside a nest—or inside the nervous system of an ant—to build a biologically meaningful computer model, Seeley says. At this point, modeling would be “a fast road to nowhere,” he contends.

Gordon’s work could make its biggest splash in other fields—computer science, robotics, artificial intelligence—where researchers hope to mimic the adaptability and division of labor shown by ant colonies. The insects exhibit what computer scientists call swarm intelligence: each individual is dimwitted, but the collective actions of the many produce apparently smart behavior, like a brain relying on millions of simple neurons. For instance, if you put an obstacle in the path of a column of foraging ants, they will find the shortest way around it.

Scientists eventually hope to give that same ability to packets of information traveling over the Internet. The packets could then route themselves around digital traffic jams and speed to their destinations. Likewise, space scientists might dispatch swarms of cheap, antlike robots to explore other planets instead of relying on a single, expensive, supersmart robot whose failure could scupper the whole mission.

As she prepares to build her “virtual nests,” Gordon is also working on a range of other projects. Next summer in New Mexico she wants to investigate the relationships between red harvesters and the other ant species living on the research site. Particularly intriguing is the harvester’s nemesis, a nervous, nocturnal ant called Aphaenogaster, which Gordon describes as “diabolical and crafty.” These ants often plug the holes of harvester nests during the night, fooling the harvesters into staying inside all day. That leaves more seeds for the devious Aphaenogaster.

This year, she is taking a sabbatical, using a Guggenheim fellowship to study at the Center for Advanced Study in the Behavioral Sciences at Stanford. She’s devoting the time to exploring her core interest—complex systems—from new perspectives. “I’m hoping to learn enough about some other complex systems, such as brains, to see whether detailed analogies [to ants] can be made.”

She cautions, though, against taking ant analogies too far. For as long as people have been watching ants, she notes, we’ve tried to draw moral lessons from them. That’s a misguided quest, in Gordon’s view. “A person with the moral qualities of an ant would be terrifyingly empty,” she points out.

Yet she does have a kind of respect for the creatures. “Their behavior clearly works; there are a lot of harvester ant colonies out there, more every year,” she observes. “I deeply admire their harvester-ant-ness, the richness of their responses to a world so alien to me.”

So take care where you step.

Mitchell Leslie of Albuquerque, N.M., writes frequently for Stanford and the journal Science.


Comments (0)

  • Be the first one to add a comment. You must log in to comment.


Your Rating
Average Rating



Be the first one to tag this!