11 negotiating the physical world
Arboreal ants scramble around the labyrinth of the canopy, spreading out along interconnected branches and vines, coursing up and down the trunks of trees. Vertigo is a human problem: most canopy-dwelling animals, ants included, are indifferent to height. What matters for them is finding the conditions and resources suited for their survival. Potential habitats occur at many levels—both on different parts of a plant and within each layer of growth in the plant community, from herbs to shrubs to shaded understory trees to tall trees with crowns in the sun, to the occasional emergent tree that sticks up above all the others.1
The ants’ choices among these parts and layers contribute to the diverse mosaic of their species in a forest. At one site in Borneo, a quarter of the ant species restricted themselves to the ground, another quarter to the understory shrubs, and another quarter to the lower canopy, leaving only a quarter to move freely through all three layers.2 That’s a remarkably low percentage of wide-ranging ants given the mobility of most workers. Low plants offer the ants living there conditions similar to those on the ground—shade, coolness, and moisture. The uppermost foliage is more fickle: in the blazing sun it is hot and dry, during periods of rain or fog it is cool and damp, and at night the temperature drops sharply. Between the extremes at top and bottom, an elevation change of a few meters up or down a tree can be equivalent to the environmental transformation we experience while traveling kilometers over the ground, say from inland mountain slopes to the seaside.
A cross-sectional view of rainforest along the Amazon River near Iquitos, Peru, showing the canopy layers, including a tall emergent tree at left, as well as vines and epiphytes. Different ants confine themselves to particular strata.
Even ecosystems with less height than a forest can offer similar variety. The overarching grasses and wildflowers in a meadow, for example, form an upper canopy. A suburban lawn can have layers of vegetation as well defined as those of a forest 90 meters high—from groundhuggers like the procumbent pearlwort and fairy flax, to midlevel scramblers such as sweet vernal, to the upright stalks of the lawn canopy giants: white clover and any of a variety of grasses.3 For a human on the ground, forest interiors provide a mild climate relative to the sun-roasted air above the trees, and a lawn’s interior offers similar conditions—for an ant, anyway.
Ants that stay within one canopy stratum, perhaps nesting and foraging their entire lives on the same tree branch, need not climb any more than ants on terra firma. However, nearly all ant species are natural climbers. Camponotus gigas carpenter ants, the biggest of all ant workers at a length of 2.8 centimeters, walk all over trees in the same Malaysian forests that weaver ants occupy. I remember huffing and puffing up one immense tree, assisted by ropes and gear, and looking over at the trunk to see a chunky Camponotus gigas major worker race ahead of me to the crown from a barrack nest at the base of the tree. Why was her ascent so effortless? When walking on flat ground, an ant burns a lot of energy relative to her mass because she has to move her little legs quickly to get anywhere; for her the added cost of a climb, compared to moving horizontally, is almost nil.4
For many people, going up a small tree can be a pleasure—at least for those of us who’ve kept a little bit of our kid selves inside. But in the tropics, even small trees harbor risks. At the Tiputini Research Station in Ecuador, I free-climbed a slim tree near the residence cabins during a much-needed break from rainy hours photographing the falling behavior of turtle ants—an effort that required scaring one worker after another off a little ledge and entailed over 6,300 clicks of my camera. I had gone up the tree because I surmised there was a turtle ant nest in it, which would provide me with a fresh supply of workers. What I found instead were workers of the giant Paraponera clavata, a species called the bullet ant for its fearsome sting. Looking at the ants lumber past my fingers on a branch next to my face, I recalled a story about one of Terry Erwin’s assistants who had overheard Terry explaining that if stung by this ant, a person should cauterize the wound with a cigarette; stung by three, the assistant had been in such pain that he’d been unable to stop burning holes in his leg. Wondering about my own pain threshold, I glanced down to see more of the rugged black workers mounting the trunk below me. Now I recalled that bullet ants nest at tree bases but forage in the canopy, where they deposit trails that guide nestmates to meals. In some cases, they arrive by the hundreds.
As a tree-climbing biologist, my anxiety that the branch bearing my weight could break is matched only by my fear of confronting bullet ants while high on a rope and unable to flee. Luckily, on that day I was neither on a rope nor high. Letting go of the trunk, I thrust out with my feet and fell the 2 meters to the ground, landing safely away from the bullet ant nest.
For humans, height is significant because we fear a fall. But falling is not the same for every being, as biologist John Haldane describes: “You can drop a mouse down a thousand-yard mine shaft and, on arriving at the bottom, it gets a slight shock and walks away. A rat is killed, a man is broken, and a horse splashes.”5 For animals larger than a mouse, there is a height above which a fall can cause harm; call it the critical-injury height. People come away from the majority of short falls with no more than a bruise, while scampering squirrels plunge meters without injury. But an ant can, in theory, fall forever without being bruised.
The distance covered, however, can have serious implications for ants. The farther a worker drops, the longer she takes to return home, and the more likely it is that she will get lost or die in enemy hands. Perhaps for this reason canopy ants are particularly good at hanging on tight. I made the acquaintance of one such species, Daceton armigerum, while exploring the Orinoco basin of Venezuela. Daceton are unmistakable: big and spiny, with lobes at the back of their head to accommodate muscles that power snap-action mandibles. These ants really knew their place within the forest strata: this colony was nesting 6 meters up a small tree, but the few that descended the trunk avoided the ground altogether, retreating the instant they touched the earth. These same ants had no aversion to walking on a humus-covered branch laden with epiphytes I placed before them. I hadn’t a clue how the Daceton workers recognized the ground, or why they found it so alarming, but for them the forest floor seemed like truly foreign soil. Given this distaste for touching down, it came as no surprise to us that after removing the tree and bisecting its hollow trunk at the village hotel, over the protestations of our maid, we found that each of the 2,342 workers could grip anything—nest, bathroom tile, ceiling—as if her feet were glued in place. I realized then how important clingy feet are to foraging and living in the treetops.
A turtle ant, Cephalotes atratus, gliding backward, toward the left in this image.
Weaver ants show no antipathy for the earth, ranging freely from treetop to ground. For them, a fall has little cost: a plummeting ant will land within her colony borders, with almost no chance of going astray. They still must avoid losing their grip when struggling with enemies and prey, which is why they resemble Daceton in having strong gripping feet. To avoid coming to physical blows, which might lead to tumbles, many arboreal ant species fight at a distance with noxious sprays.6
One way or another, ants do fall, and in such numbers that they can amount to an ant rain. Ground-dwelling European wood ants, Formica aquilonia, which forage in the trees, plummet from the canopy by the millions each day. Formica rain harder, so to speak, when near foraging birds. Some are knocked off branches, but others jump to avoid bird pecks. Still, the ant rain continues even when there are no birds. A portion of these ants lose their grip by accident, but some might fall simply to save time on the commute back to their nests.7
Whereas ants who don’t live in the trees tend to tumble willy-nilly and hit the ground blindly, canopy species frequently seem to be able to control their falls. Jack Longino of Evergreen State College and I have spent some time dangling from ropes, where we’ve often contributed to ant rain by knocking various species of ants off branches, for the most part unintentionally. More often than not, it seemed to us, the ants would land back on the tree, as if they could control their flight path and hit a target. We couldn’t understand how they did it.
Stephen Yanoviak, then at the University of Oklahoma, noticed the same thing and set out to prove that certain ants from Peru and Panama indeed can glide.8 The species he concentrated on was Cephalotes atratus, a slate-black “turtle ant” with a flattened body. High-speed videos proved that when dislodged from a tree, a turtle ant stretches out her body and limbs and aligns herself with respect to the ground so that she doesn’t turn head over heels. Detecting a tree trunk by its relative brightness against the dark greenery, she twists in the air to point her abdomen in that direction, glides backward at a steep angle—a behavior that I was eventually able to capture for this species with my camera in Tiputini, Ecuador—and grabs the trunk on impact.9
Other ant species make a tight spiral as they fall, directing their bodies with apparent intention, much like a parachutist who aims well enough to strike the earth at a good spot and on his feet. In a rainforest, numerous leaves lie between a plummeting ant and the ground. I believe that if she can slow her descent while keeping her clingy feet oriented downward, a worker can greatly improve her chances of landing securely on one of these leaves rather than bouncing off, as she’d surely do if she were tumbling head over tarsi.
The method employed by a falling worker depends on where she lives and the dangers she faces. Weaver ants neither glide nor spiral down, but rather plummet head over heels, a reflection of how little a fall matters to them, whereas Daceton are virtuoso gliders. Both the turtle ant and Daceton nest in tree trunks or thick branches. When one of these ants falls, a trunk is likely to be in range, and gliding to it is the obvious choice. For species that nest farther out among the twigs it makes little sense to aim for a trunk that may be too far away to see, let alone reach by gliding. Foliage is a sensible target, and parachuting in a tight spiral is the way to make a firm landing.
TRAVELING IN THE CANOPY
What convoluted territories arboreal ants inhabit! The navigational problem faced by a small ant is that a tree for her is a highly warped surface, one that is much more complicated than the surface of the earth is for us. She can usually monitor her movements up and down by gauging the influence of gravity on her body segments, but these gravitational effects can be masked by the movement of a plant in a breeze.10 Because she may often have no idea which way is up, a worker in the canopy does not experience the geometry of the world the way we do. She can walk in one direction and find herself back at her starting point (she circled a trunk or branch). If she makes a ninety-degree turn, she may either reach the end of the world (a branch tip) or be lost forever (having walked down the trunk and onto the ground).
It helps that individual trees have some common features, such as a limited number of branching patterns—compare the alternate branching rhythms of an oak to the terminal leaf clumps of palm, for instance.11 Unlike a rat forced to navigate a psychologist’s maze that has been constructed with no thought to the geometry of nature, arboreal animals can use a tree’s predictable structure as a navigation aid.
Ants exploit many aspects of plant architecture. On trunks, columns of workers often follow grooves or edges to orient themselves upward. When a worker reaches a horizontal leaf splay, she tends to stay on its upper surface, where she is less likely to be knocked off. She can survey a leaf by moving along its edge to trace its outline, deviating if she chooses to explore the leaf’s center or underside before going back to the margin. When she departs a leaf and encounters a branching point among the twigs, her best bet is to always turn onto the next stem in a consistent direction—say, to her left. By tracing leaf outlines and being consistent in her choices at forks, she is able to move through a leaf spray without revisiting the same leaf twice and without ever having to mark a route or memorize the terrain.12 This technique enables an ant to examine a plant more efficiently than is possible when she explores the ground. That is true even when the ground is flat and bare, which is rare: ground-dwelling ants navigate through jumbled decaying matter and plant parts, geometries far more haphazard than trees. For weaver ants, the solution is to commute on the ground as they do in the canopy, following crestlines offered by exposed roots or fallen sticks.
Once an ant begins to range widely from one tree to the next, the messiness of canopy topography forces her to rely on a variety of orientation cues. Leafcutter ants, for example, measure thermal radiation to locate the sun-warmed foliage they prefer to cut.13 Visual cues may also be valuable. A weaver ant can choose a course at a particular angle relative to the sun or moon; if the sky is shaded, she uses a less accurate, internal magnetic compass.14 Some ants create maps by taking mental snapshots of the greenery against the sky.15 Although it may be relatively easy to use these snapshots to navigate on flat ground, using them within the trees must require an overwhelming feat of insect memory.
A good memory can be essential to canopy survival. The weaver ants Ed Wilson raised on a small citrus tree in his office when I was his graduate student became excited when a novel object such as another tree was placed nearby, gathering on a twig in an attempt to reach it. Apparently they remembered enough of their surroundings to recognize this change. We might expect the workers in a large colony to keep to the small portion of the territory that they know well, and in fact weaver ants on border patrol don’t move around a lot.16
Age and experience can play a big role when ants explore farther afield. The chemical trails of bullet ants, for example, often cross one another among the interdigitating branches and vines of the canopy, and the workers can distinguish the scent routes laid by different nestmates to different destinations.17 These ants also have exceptionally good eyesight and keep track of their whereabouts by eventually memorizing the location of such landmarks as branches.18 Novice workers follow the trails, while their more experienced compatriots come to navigate by the landmarks almost exclusively.19
In Queensland, weaver ants—known in Australia as green tree ants because of their coloration—form a chain to connect branches during a foraging expedition. In most situations the chain is many ants thick.
It’s not easy even for arboreal species to range beyond one tree. Rainforest crowns are separated by open spaces and seldom intermix. Vines offer shortcuts as well as an abundance of honeydew-producing insects, explaining why weaver ants prosper at forest margins where the canopy is cluttered with such connections. Adjacent tall trees lacking vines tend not to be occupied by the same weaver ant colony. However, in these situations weaver ants can create their own shortcuts. I saw this ten years ago during a stay on the north coast of Papua New Guinea. I was drying myself off after snorkeling when I noticed, stretched between the branches of two citrus trees a few meters above my head, a chain of weaver ants 6 centimeters long. I broke the chain with a finger to see what would happen. Ants accumulated at the site, climbing on one another in the direction of the neighboring tree to form a fingerlike mass jutting into the air. After an hour, the wind rocked two branches close enough for the workers at the end of the “fingertip” to grasp a twig on the opposite tree. As the trees swung apart, the chain pulled taunt and the bridge from one tree to the next was renewed.20 The ants had created an easy route to avoid the long march down to the ground and then over to and up the second tree. The next morning, the ants were still there, strung together and straining under the weight of a small cicada being carried across their bodies by a team of their nestmates. Being part of such a skywalk has my vote as the worst job in an ant colony.
Much of my time in the tropics has been spent in one of two uncomfortable situations: wedged between branches high in a tree or traveling along the ground in a downpour so warm and torrential that I felt like I was suffocating. Luckily, these are superb times to ant-watch, with the goal of understanding an ant’s physical world. There’s a rule of thumb among biologists: living things in temperate zones suffer more from physical events such as cold spells, while creatures in the tropics are more vulnerable to biological threats, such as predators or competitors. Flooding is an exception to the rule in that it is a physical problem for ants no matter where in the world they live.
Flooding can be serious even for canopy ants. Southeast Asian bamboo-dwelling Cataulacus muticus ants use their helmet-shaped heads as rain shields to block the holes that allow access to their nests. If a chamber still floods, there’s an unusual backup plan: hundreds of workers drink their fill, climb outside, lift their abdomens high, and communally pee. The removal of one milliliter of liquid requires the ants to relieve themselves a total of 1,515 times.21 Other species bail out their nests by spitting the water away with an audible click of their mandibles or smearing it on the ground outside their nest.22
Weaver ants’ leaf nests shed water, but having been in a tent that collapsed from the force of a cyclone, I know firsthand what a wild ride they experience during heavy storms.23 Meanwhile, any ants caught outside take a beating. When a shower begins, the workers that had been climbing to their nests turn around and walk in reverse, with their legs spread wide, giving their adhesive toe pads a better grip. If the rain turns torrential, they find shelter under a branch or, failing that, aggregate in clumps that grow in size and compactness until as many as a hundred ants are huddled together, limbs entwined, holding tight to the bark to keep from being washed away. As soon as the rain slackens, they disentangle themselves and scurry home, head forward.24
Ants on the ground have it worse. Within a marauder ant nest, things get soggy as water swirls through passageways, though the entrances are often higher than the surrounding ground, which helps keep water out. Marauder ant workers caught in the rain will, if they’re lucky, make it back to the nest or find a sunken or underground section of a trunk trail to retire to. As a last resort, the ants withdraw beneath the leaf litter, though this is likely to become submerged as puddles form. Once I observed a leaf break free of the bottom, dislodged by a bubble that then burst to the surface, with soused ants swirling into view. Apparently, the workers had found safe haven in an air pocket below the leaf. The unlucky ones drowned.
Once the rain stops and the puddles soak into the earth, the marauder ants emerge from their hideaways to reorganize into columns. Heavy downpours obliterate physical evidence of a trail’s existence and must wash some of the pheromones away, too. A trail reemerges as the workers locate each other, linking up in a route that usually manages to follow the old one closely. Workers that fail to find their nestmates become lost and die.
The red imported fire ant of the American South—originally from floodplain habitats in Argentina—survives deluges by forming a raft of thousands of workers and larvae, with the brood and queen nestled in its center.25 University of South Florida professor Deby Cassill has observed such a pontoon. The workers reach jointly for land in much the same way that a finger of weaver ants reaches through the air for another tree: “I have noticed,” she told me,
that while pupae sink and are often lost during a flood, the larvae float, especially sexual larvae, which are full of bubbles of oil or maybe gas—a product of digestion in a closed one-way digestive tract. If you look at the raft from underneath, you see the larvae being used as inner tubes, held together by the grappling-hooked feet of the workers. On top of the raft, the workers around the edge reach out with their forelegs to grab anything that floats or is anchored. As they reach, other ants walk on top of them, grapple onto them and stretch out over them. So the raft begins to look like an amoeba, with arms of ants extending from the edge in little fingers.
She went on to recount how one of the fingers eventually latches onto a branch or piece of grass, which is followed by a rush to the shore.
In the Amazon basin, seasonal flooding caused by runoff from the Andes drives terrestrial invertebrates into the trees, which serve as a commodious version of Noah’s Ark. Entire ant colonies are among the menagerie ascending the tree trunks, although it’s unclear how all of them find makeshift living space there.26 As conservationist Michael Goulding writes,
At the beginning of the rainy season . . . soil arthropods begin to migrate upwards to the trunk and canopy layers with spiders, millipedes, and centipedes being especially common. Most arthropod groups appear to migrate before the actual inundation starts. . . . [Others,] however, wait it out, and only leave the forest floor when it is flooded. Sow bugs (tiny crustaceans) and small spiders are among these adamant groups. Spiders especially, but also predaceous ants, form a veritable gauntlet that upward-moving, flood-fleeing invertebrates must run.27
Probably most ants can swim to some degree, though to my eye, marauder ants appear to do little more than flail wildly. As the Amazon waters rise, leafcutter workers can walk on the water surface to nearby tree trunks or swim to them when flooding becomes severe. If a worker misses a trunk, she stops swimming and floats along until she passes near another.28 The common carpenter ant of the eastern United States is equally good at swimming. To generate thrust, a worker moves her forelegs in the same manner as she does when walking, while employing her middle and hind legs as a rudder for making turns. For another carpenter ant, the giant Camponotus gigas of Malaysia, swimming, like climbing, is no big deal; instead of detouring, workers paddle across any puddles in their path.29 Indeed, a Camponotus gigas should find it easier to swim than a marauder ant, because water offers more resistance to a smaller individual. However, the marauder will carry more air down with her, proportional to her size, which she can use for breathing and to make herself more buoyant for her slow haul back to shore.
A Camponotus schmitzi worker free-diving into the digestive fluids of a pitcher plant in Brunei, Borneo, where it will retrieve the corpse of a cricket.
In the mangrove swamps of northern Australia lives an ant that swims as a matter of course. Nests of the spiny ant Polyrhachis sokolova can remain underwater for several hours at high tide. As the waters begin to rise, the workers swim on the ocean surface to reach the raised entrance cones in the mud, rowing with their front two pairs of legs and using the back pair as a rudder. Once inundated, the sandy cones collapse, plugging the colony safely inside. If an ant doesn’t get back to the nest in time, she awaits the return of low tide on the trunk of a nearby mangrove tree. When the waters recede, the nests are opened by the ants sealed underground, which then walk out to hunt small crustaceans on the mud flats.30
One plant-dwelling carpenter ant species has incorporated swimming into its foraging routine, in a behavior so extraordinary I had to see it to believe it.31 To do so, I returned to Brunei, home of the exploding cylindricus carpenter ants, and drove an hour west of the capital, crossing most of the breadth of this tiny country to a remnant patch of red meranti, an endangered timber tree with a long, pale trunk. There I found Camponotus schmitzi workers crawling on pitcher plants that grow as vines at the base of the meranti.
A pitcher plant is not ordinarily a healthy place for an ant, since these plants are carnivorous. A cup that grows from the twisted tendril at the end of each of their modified leaves holds a liquid into which insects tumble and drown after “aquaplaning” over the pitcher’s slick rim much as a man slips on a banana peel.32 The plant secretes digestive enzymes into this liquid that break down the corpses and help the pitcher absorb their nutrients. Ants are the plant’s most common meal, except for the resident species of carpenter ant, Camponotus schmitzi, which nests in the pitchers’ tendrils and takes dips in the liquid, emerging alive and well.
That afternoon, I watched ants dive into the cups for a swim, staying underwater for up to thirty seconds. At the floor of one pool, two workers tugged at the corpse of a cricket, dragging it up through the water meniscus—a feat in itself, given how difficult it is for a small being to break the surface tension in a body of water. Then the pair carried the body up the pitcher walls, an equally tough job because the surface is slippery, thanks to a flaky wax that helps the pitchers entrap their prey. Slowly, the ants dragged the cricket to the underside of the pitcher rim, where a dozen other workers gathered to eat it.
What looks like theft turns out to serve the plant. By working in twos or threes, the little divers retrieve insect corpses several times their weight. These bulk items can’t be tidily digested by the plant and so tend to putrefy. Liquid fouled with ammonia and sullied by organic matter gives the pitcher plant the equivalent of acid indigestion and causes the pitcher to rot. The ants therefore aid the plant by removing large prey, but they also feed it: as my ant workers ate their cricket at the plant’s rim, small chunks of the insect dropped back into the liquid below, to be absorbed by the pitcher plant. For Camponotus schmitzi the pitcher is a first-rate “ant plant,” providing for its residents’ every need: housing and meat, and even sweets, in the form of the nectar at the rim of the pitcher that also attracts its hapless prey.
DOES SIZE MATTER?
Whether walking, swimming, climbing, or falling, an ant’s diminutive size influences how she travels through her world. Although we think of all ants as small, they vary in size several thousandfold. The average species has workers a little less than 3 millimeters (an eighth of an inch) long—smaller than a weaver ant and about the size of a marauder ant minor worker. But ants at the small end of the spectrum, such as Carebara atoma, the “atom ant,” are truly Lilliputian. I once dislodged a flake of bark from a tree in Singapore, only to expose four hundred yellow specks: an entire colony of its close cousin Carebara overbecki. The minor workers were almost the size of an atoma, their oval heads about as small as a single-celled paramecium. The slightly larger soldiers have elongated heads with two little horns.33
The ant worker at the other end of the scale, the major of the carpenter ant Camponotus gigas, is nowhere near the car size of the ants that terrorized Los Angeles in the 1950s cult film Them! At a little more than an inch long, she is indeed only fair-to-middling in size among insects and falls far below the world record holder for an adult insect, a female giant weta cricket I collected on Little Barrier Island in New Zealand (it weighed 71 grams, three times as much as a lab mouse). One scientist observed that ant species with bigger workers tend to show a greater number of behaviors, and he proposed this might be because of their larger brains.34 Still, Camponotus gigas workers don’t strike me as being especially quick-witted, and indeed there are many physical advantages to staying relatively small. Although the ant’s little body loses heat and water more easily than yours or mine and overheats more swiftly in the sun, it also circulates nutrients without as complex a cardiovascular system. As we’ve seen, an ant’s size enables her to climb almost effortlessly and fall without the possibility of breaking a leg. Ants can float or swim when caught in a downpour, and survive long periods of immersion thanks to their sluggish metabolisms. (Yet weaver ants drown inexplicably fast, making it easy for Cambodians to collect them in water and pull out the bodies later to add as spice to a meal.)
But looked at another way, ants aren’t small at all. A leopard may impress with her bulk and power, but compared to the ant, she is a minor part of the forest in which she lives—measured in terms of both her ecological impact and her size. Ants have, in effect, two body sizes: the individual’s and the colony’s. To understand this basic truth, I use a mental exercise I learned as a graduate student studying the marauder ant. First, I follow an individual ant. Then I take in several ants collectively, a group of workers busy at a task. Finally, I liberate my imagination from what is directly before me, emulating German chemist August Kekulé, who discovered the beautiful structure of the benzene molecule in a dream. Allowing my reverie to expand beyond what is visible, I contemplate the functioning of the whole: all the ants, in the nest and out, with the workers integrated like the cells of a human body into a superorganism.
This is more than mental gymnastics. By living socially, ants break through the glass ceiling imposed by their exoskeleton. At nearly 40 kilograms, a large driver ant colony is the size of an eleven-year-old boy. However, this particular young man would be a kind of superhero, one who can disassemble, such that his hands can stop a crime while his head commutes to the office to write up the news report—both Superman and Clark Kent at the same moment, an analogy that is particularly apt for ant species whose workers spread out widely. Even an atom ant colony, which might fit within the head of a Camponotus gigas worker, is a superhero in miniature.
A colony is an organism divided, with no loss of integrity. Its body spreads over space in pieces that give it a multitude of eyes and brains with which to glean nutrients, energy, and information. It does this with a microscopic attention to detail that no unified body can match. It’s more flexible than an organism, too: the superorganism counterparts to tissues and organs range from transport teams to nurseries and are easily assembled or taken apart or shifted to a different function. Whereas a human vascular system has well-established roles, its colony analogue, the trail system, is flexible; it can serve as a snare for food and later be co-opted for migrations or a fight.
This fragmentation helps a collective of ants to succeed when a single big vertebrate would fail. The workers in a weaver ant colony, with a combined weight, at 14 kilograms, of a young leopard, can disperse among leaf nests on many frail branches—or, in other species with big colonies, fill up cracks, crevices, and galleries in wood—and thereby live where no hefty vertebrate could. Furthermore, most of the food available in nature is present in packets too small for a large animal to glean. A young leopard or a man would starve trying to gather the tidbits that make up the diet of a large ant colony, and neither is muscular enough to carry as much food as all those ants can move collectively. A whale trawling for zooplankton is the only vertebrate creature that scoops up as many bits of prey; indeed, a baleen whale is the marine equivalent of an ant colony.35 To accomplish this task as a group, bulky individual ants (even ones of weta cricket size) would be at a disadvantage. That’s why so many plants have evolved to support ants, but not aggressive mammals or birds, as their guardians. Only ants can scour a plant’s surface relentlessly enough to weed out its enemies, large and small.36
Their scouring behavior illustrates the repetition and fast tempo that are the hallmarks of large ant societies. We saw this with the swarms of marauder ants that crisscross the terrain within a raid, rooting out prey (and weaver ants do much the same in much looser bands). An advantage to having many do the same thing at once is that if one individual fails to finish a task, whether subduing prey or building a trail, another will do it.37 Also, the frenetic workers in large societies often make mistakes; close inspection may reveal one ant going the wrong way or leaving building material at an inappropriate site. Such a blunder might be lethal for a solitary creature that has but one chance to do a job right; the same may be true for ant nests with few individuals, in which workers carry out every move with meticulous care. But in a large society, differences in performance assure that often enough a task is done to perfection. Even if lapses or errors occur, they are quickly corrected by another individual. In fact, with sufficient redundancy, variations in performance can lead to useful novelty and innovation, as when an ant on a busy recruitment trail overshoots the intended prey and, while she is lost, happens on another; or when a friendly competition for goods and services brings a kind of market economy to a nest.
The redundancy of worker actions gives a superorganism other survival advantages as well. While a human life ends if a wound destroys the brain or heart, the functions of brain and heart are spread throughout a colony, making it harder to damage. To bring the comparison to the level of the society, we humans have erected increasingly elaborate top-down hierarchies and centralized systems of control to deal with the disasters, from plagues to terrorism, that so easily disrupt our modern nations.38 The lesson from nature, however, is that the war on terror will never end: all living things fight back against enemies (parasites, predators, and competitors) in a continual arms race in which new defenses emerge but the dangers never disappear because the adversary always evolves a counterstrategy. Under such circumstances, it doesn’t pay to consolidate power; better to have redundant operations with few or no established chains of command, as ants do.39 Because they have no central command center that can malfunction or be crippled or manipulated by outside forces—indeed, no established leaders or unique individuals—the destruction of part of the population will never bring down the whole.40 Weaver ants apply this redundancy even to their nests by having multiple leaf tents instead of a vulnerable central domicile. With the exception of reproduction (most colonies succumb to the death of the queen), this safety net permeates all aspects of ant social existence.41
Is there an ideal size for individual workers within the superorganism? The answer is unclear—as it is for the size of the cells in the body, for that matter. One weaver ant major worker can stretch across the nail of a pinkie finger. Among ants, that’s pretty large. Weaver ants compete with other dominant ants that have much smaller workers. But imagine a mature weaver ant colony that, instead of being polymorphic, contained only minor workers, and instead of half a million ants, held several million. The colony consisting only of minors would burn more energy at rest than the original colony or a colony of only large workers of the same total weight—perhaps an economic disadvantage; and no minor worker could hold prey as large as one major can restrain for long enough for reinforcements to arrive. But the foragers would exchange information at a higher rate and be more effective at rooting around in locations previously hidden to them; and the greater number of individuals would be more effective at ganging up against the competition.
Variation in worker size is associated with a division of labor, and the redundancy afforded by large ant societies helps make specialization pay off, as it does for large human groups and complex organisms. With humans in, say, a small military squadron, the loss of the one person who knows how to radio headquarters could be devastating; in small groups it therefore pays for everyone to be a generalist, with overlapping knowledge and skills. Larger divisions can afford to include more specialists, among them helicopter pilots, tank drivers, and snipers. Large ant societies can similarly produce more specialists. Consider the outrageous polymorphism of the marauder ant: by having workers that span the size distribution of many of their competitors combined, they may be able to outperform them all the more effectively.
We don’t know why, given a certain outlay of resources, one ant species produces only large workers, another produces only small ones, and a third—the weaver ant—produces a mixed population, biased toward the majors. Yet so omnipresent are the major workers, so complete their concerted action when they sense a person, that I have often felt, on walks in Africa or Asia, as if some predator were spying on me from the trees—only then to hear on all sides a muted drumming of alarm, similar to the sound of peas dropping onto a plate, as one study puts it.42 That is the collective sound of a superorganism, generated by the crowds of agitated weaver ants striking their leaf roosts.