In a pasture near Botucatu, Brazil, two cows were staring at me with heat-addled eyes, when all hell broke loose. Luiz, a laborer we had hired to help us, gave a shout of pain as part of the trench he was digging in collapsed. In front of him was a gash wide enough to hold a treasure chest, from which spilled not gold but a porridge-like material. Despite Luiz’s screams, my heart started beating with the same excitement a prospector must have felt when he struck a vein of gold. I ran over and lowered myself into the trench.
We were digging for leaf-cutting ants. These fall into two genera, Acromyrmex and Atta, with a total of thirty-nine species, all from the New World.1Atta, the most impressive and ecologically important of these ants and my focus in this chapter, are most prevalent in the tropics, though Atta texana range through east Texas and west Louisiana, and Atta mexicana cross into Arizona. Their popular name derives from the medium-sized members’ habit of cutting foliage. They hold the pieces aloft like little green parasols, then stream across the ground to their immense lairs, where they use the plant material as a substrate for underground farming.
In four ten-hour days, our ten-person team—led by Virgilio Pereira da Silva of São Paulo State University—had so far only scratched the surface of a labyrinth of chambers that seemed to go on forever. It was thrilling to think we might finally be reaching the heart of the nest. I now stood in one of two 7-meter-long corridors we had dug, both deeper than we were tall. Everywhere along their walls were bisected galleries. A leafcutter nest can extend 7 meters into the earth and contain nearly eight thousand chambers. The biggest hold the ant’s trash, buried—with greater thoroughness than humans use in handling nuclear waste—as deep as the ants can go, sometimes as far as the water table. In this colony the largest cavities were still below us. But we were getting close. The gash Luiz had just opened revealed several kilograms of this refuse. Passing my hand through the loose stuff, I detected first the heat of decay, then motion. In my palm wriggled browsing beetle and fly larvae, a moth pupa, and several cockroaches—species that speed the breakdown of the material.
My excitement wasn’t much consolation for Luiz. His back was covered with leafcutter soldiers the size of horseflies, which had latched onto his flesh with scalpel-sharp mandibles, and his shoulder was bleeding where a soldier was busy slicing an especially deep curved groove.2 I indicated to Luiz that he should protect himself by wearing a shirt, but my own defenseless hands were marked with similar crescent moons, painful as paper cuts, as if I’d been clawing through brambles. More soldiers poured out of a hole in the floor of the trench and swarmed up my legs as I helped two of Luiz’s friends pull the ants from his back. Then everyone picked up their picks and shovels and went back to work.
Leafcutter soldiers have few duties. They have been known to use their offensive skills to dice up tough fruit, but mostly they protect the nest and its immediate environs from army ants, hungry armadillos (whose powerful forearms help them tunnel easily into the heart of a nest), and curious entomologists and their unfortunate assistants.3 The damage they had done to Luiz’s back and my hands reminded me of Lophomyrmex bedoti, an otherwise unremarkable ant with jaws almost as fine-toothed as saw blades. In Indonesia, I had seen three workers of this species shear off both antennae, four legs, and the tips of two body spines of a big-headed ant with grisly ease, in less than thirty seconds.4
I would have guessed that leafcutter jaws work with an equal effortlessness, but leaf carving is an arduous activity for these ants, equivalent to the cost of flying in other animals.5 Follow the leafcutters to the source of the ants’ parasols, and you will find the ants on vegetation, cutting arcs like the ones chiseled into my fingers. Rather than using her mandibles like saws or scissors, a leafcutter sticks the terminal tooth of one jaw into the leaf to fix its position as she pushes her other jaw against the leaf edge. She then forces the second jaw into the tissue with a rocking motion, the way people use a lever-type can opener. Leafcutter jaws are also like a can opener in that they get some of their strength and rigidity from metal, having a zinc content of 30 to 40 percent.6
While cutting, the worker herself acts like a geometer’s compass: she anchors her back legs at the leaf margin and moves in an arc around that point, adjusting the size of each fragment by flexing her legs or head to different degrees. By such fine-tuning, she slices off fragments that she (or other ants) can carry: smaller pieces from thicker leaves and larger pieces from thinner ones.7 She can also be precise in her choice of leaf: workers tend to end up on the foliage they cut best. This is often based on their size, which is quite diverse in leafcutter workers, even among those who specialize in cutting leaves. The smallest cutters abandon foliage too tough for them to handle, while larger ones tend to depart from soft foliage after being pushed aside by bustling smaller ants. There’s even some evidence that larger cutters are drawn to recruitment trails that lead to the sturdier foliage.8
Unlike marauders and army ants, leafcutters always move their burdens individually and never need help.9 Because leaves are flat, large pieces have a small mass that a single ant can easily heft, and larger individuals slice and haul larger loads.10 But leafcutters take fragments just a few times their body weight at most. They are capable of hefting heavier fragments, so this may seem inefficient, but small pieces reduce congestion on the trails and inside the nest, speeding up the processing of foliage overall.11 Additionally, because workers carry the fragments vertically and high above their heads, on steep slopes the ants can become unbalanced, causing them to slow down, flip over, and even fall. The ants seem to anticipate when the route home will be uphill, and cut smaller parcels.12
HOW DOES YOUR GARDEN GROW?
Leafcutters are almost unique among ants in their total dependence on vegetable nutrition. True, many other species rely to some degree on plant matter, notably in the canopy, but vegetation contains meager protein. Hence, treetop ants like the weaver ant are omnivores, the workers using the energy they get from carbohydrate-rich plant food to procure animal flesh for their protein-greedy larvae. (Only the dwellers of a few ant plants can afford to be strict vegetarians; their plants secrete protein-rich food bodies in payment for the ants’ protective services.) Leafcutters have a similar strategy.13 The adult workers drink sap from leaf fragments (their primary energy source), energizing themselves to process the foliage that serves as compost for their larvae’s sole food: a protein-rich fungus that is related to the button mushroom sold in supermarkets. Leafcutter colonies have come to depend on the fungus so much that even the adult ants’ digestive tracts lack key enzymes for breaking down some proteins.14 The fungus does that for them, and also removes from the plant tissues any insecticides that would stop most herbivores in their tracks.15
Most people know fungi as mushrooms, but mushrooms are the substantial yet fleeting reproductive outgrowths of microscopic strands called hyphae that spread in a latticework to infiltrate and ingest soil, decomposing matter, even rock. In the leafcutter nest, this latticework spreads through the foliage that the workers have mulched in order to free its contents for absorption, a necessary step given that the fungus likely can’t digest cellulose and only takes nutrients in solution.16 The hyphae and its vegetable substrate fill the majority of leafcutter nest chambers with what is called a fungus garden, a mass of featherweight, fissured gray matter that looks like a human brain and can reach a similar volume. A garden is given its cerebral shape by workers that continually add fresh leaf matter to its top and sides while dismantling and disposing of the bottom, older half.
Leafcutters grow and harvest their fungi using farming techniques no less complex than ours. Along with the fungus-growing termites I saw attacked by driver ants in Africa, they are among the only animals besides humans that can be considered agricultural.17 The invention of agriculture has enabled societies of humans and leafcutters, which were farming long before people, to support massive populations. These massive populations then give rise to massive structures—for shelter, food rearing, and so on. The most enormous nest I have come across was in the dense forests of the Kaw Mountains of French Guiana, with soil mounds rising chest high over an area 14 meters wide and about 160 square meters total. Trails initiating from the nest’s far corners led into the forest in each cardinal direction. Scaled to human size, the space occupied by such a nest would exceed the dimensions of the Empire State Building. Such a colony might easily contain several million workers.
A small fungus garden of Atta colombica in Paraguay: the white fuzz is fungus, which is tended by the workers. Hiding in the nooks are winged queens.
A leafcutter colony’s chambers and tunnels can require the excavation of 40 tons of soil, as they must house not only queen, brood, and workers, which even in the millions occupy only a tiny fraction of the space, but also fungus gardens in the hundreds or even thousands. The garden chambers are distributed along tunnels in a pattern that can resemble grapes on a stem—with the garden-containing “grapes” the size of soccer balls and “stems” as wide as a child’s arm, which must give the jostling leaf-bearing ants elbow room aplenty.18
Cramming gardens and millions of ants together underground produces air pollution, and too much heat or too little oxygen will slow garden growth. In the Kaw Mountains, I took time off from swatting mosquitoes to peer into a nest entrance that thrust from the earth like a volcanic cone. The metabolic heat of fungus and ants struck my face like the exhalation of a great bull. Illuminated by my headlight, the smooth throat, over 7 centimeters wide, gracefully curved out of view a meter or so down. This vent was near the center of the nest, where the population is densest and the heat of metabolism therefore highest; humid air escapes through such openings, to be replaced by fresh, cooler air drawn through perimeter entrances. In open habitats, wind striking these turrets could be the principal source of air conditioning.19 For these reasons colonies can have a thousand entryways; those not being used as ant thoroughfares can be opened and closed to regulate the conditions below.
A colony this big is comparable in many ways to a cow or a deer. The ant population weighs 15 to 20 kilograms, as much as a newborn calf—or an adult red brocket, a Latin American deer that lives in the same forests as many leafcutters. Leafcutter nests consume as much vegetable matter in a year as does one red brocket: up to 280 kilograms, enough leaves to blanket a soccer field.20 As we shall see, the assemblage of ants in a colony processes forage in much the same way as a cow does, from chewing the raw material to excreting the remnants. The gardens are the equivalent of the cow’s rumen—but whereas the rumen makes use of a slurry of bacteria, protozoa, and fungi to extract the proteins and fatty acids a cow needs, a garden requires only one fungus species to process foliage into a complete ant chow.21 Both ant and cow find and prepare plant matter for their microbes, which they house under ideal conditions; the ants will relocate gardens if the temperature or humidity falls outside a suitable range.22 Just as algae and fungi have combined to form an organism known as a lichen, and the gut flora have become an essential part of a cow or a deer, the garden fungus has been integrated into the leafcutter superorganism.
INDUSTRIAL FARMING AND TRANSPORT
Farming requires a diverse skill set in ants, as it does in humans. Today, humans farming on a large scale use tools and machinery to handle different steps in the process, but in ants, different skills reside in different workers, and as we have seen, polymorphism plays an important role in this. The biggest leafcutter colonies are extraordinarily polymorphic, with the largest soldier having two hundred times the mass of a small worker.23
Ant colonies have been likened to a factory in a fortress.24 I find the metaphor particularly apt for leafcutters. Their multiple-step procedure, in which all get involved, dwarfs the two-step process by which a marauder-ant media worker extracts a seed from a grass stalk and a minor carries it away. The leafcutter workforce is self-directed, adjusting to the local requirements of colony and fungi without the oversight of any foremen. In business terms, it has the flat organizational structure adopted by corporations from Hewlett-Packard to IKEA, with the absence of middle management enhancing cost effectiveness and the organization’s responsiveness to rapid shifts in needs.25 Most leafcutter activities are accomplished with little communication: as is done in any well-run assembly line, the gardeners simply do the task that comes before them.
A leafcutter factory might have been the envy of Henry Ford: different workers collect, transport, and mince foliage, apply it to a garden, and eject its decayed remnants in an orchestrated flow of material from environment to nest and back out again. Many steps are managed by ants in a narrow range of sizes.26 Mid-sized workers cut the foliage, carry it into the nest, and drop it onto the garden surface, where, as the production line unfolds, ever smaller ants accomplish more delicate tasks. Workers with heads about 1.6 millimeters wide shred the greens into scraps. Slightly smaller ants further masticate the chunks, now discolored from abuse, into a moist pulp. Still smaller ants, using their forelegs, implant the pulp into the garden. Tiny ants with heads a millimeter wide lick the pulp and seed it with tufts of fungus from established parts of the garden, like horticulturists using cuttings from a vine to begin a new crop of grapes. The smallest workers of all, with head widths of 0.8 millimeter, reach into the garden’s recesses to remove weedy species and contaminants that include bacteria, yeasts, and spores.27
This small leafcutter worker from a Panamanian colony of Atta sexdens is taking a tuft of white fungus to “plant” in fresh leaf mulch.
Much like vintners trimming back the branches of grape vines to maximize their yield, minor workers prune the garden surface, stimulating more edible fungus growth.28 The gardens’ brainlike fissures dramatically increase the surface area from which minors can harvest meals.29 The small workers seem to take on the primary role in distributing food. They drink sap while processing leaves and nibble on the fungus while tending the garden recesses. Then they feed their nestmates, either by regurgitation or by handing them edible wads, which they also give to the larvae scattered over the garden.
The leafcutter ant Atta vollenweideri plays an important ecological role in Paraguay, where the Caranday palms and mesquites of Chaco savannas sprout from the fertile soils of dying nests.
All along this conveyor belt, the workers defecate on the leaves. Their feces, like the manure we use in our gardens, contain ammonia and amino acids that promote garden growth. Leafcutter excrement also includes enzymes from the fungi they have eaten, which pass through their digestive systems intact and speed the breakdown of fresh substrate, helping each new tuft get better settled. Coddled and cared for, in a day the fresh fungus has sprouted what look like microbe-sized masses of cotton candy. The masses consist of swollen hyphae tips, configured to lie in easy reach of hungry ants. As seen under the microscope, they are arranged in grapelike clusters, much as the fungus garden chambers are along their underground runways, but at one-thousandth the size. Found in no other fungi and with no purpose other than to be eaten, the swellings reveal that the fungus has been selected over long periods of time by the ants to serve as food, just as plump grapes and rosy apples show generations of cultivation by human hands.
As in human industries, such large-scale operations require the support of extensive transport and distribution networks. On average, each leafcutter colony maintains a trail system 267 meters long at any given time, which requires the completion of 2.7 kilometers of roads over the course of a year. That much construction requires 11,000 ant-days of labor, during which the workers expend the energy attained from eight thousand leaf burdens. That sounds like a lot, but because a leafcutter workforce is larger than the human population that was employed to build the pyramids of Egypt, it takes a colony less than a day to fetch enough foliage to fuel a year’s worth of road building.30
I came across some prime leafcutter trail systems in Paraguay in the early 1990s while driving with a companion through the Gran Chaco, an alternatively swampy and scorchingly dry savanna region. The thermometer read 125 degrees Fahrenheit, and the heat was overpowering—literally: on our second day, my friend was catatonic from exposure, staring blankly into space for half an hour before I could rouse him. We had come in search of Atta vollenweideri, which can unearth more cubic footage than that boasted by an average New York apartment.31 The nests, with their excavated discs of pale sand, several meters across and visible to the horizon, each radiated trunk trails as wide as a human foot and more than 70 meters long, leading to the grass this species prefers, which it cuts into long, linear segments. Days later, I had a chance to view the area from a low-flying airplane. The leafcutter communities resembled highway maps of human urban centers; in fact, some leafcutter species reportedly have beltways encircling their metropolises.
The leafcutter transportation system comprises durable trunk routes with weaker side trails near their far end from which the ants spread out to forage.32 In canopy-foraging species, each side route lasts for a few days, during which time it grows longer because, as ants harvest foliage, a lack of space for late arrivals forces them to move out to explore other, interconnected vines and branches.33 These temporary outgrowths often cause the layout of paths for a colony to resemble a two-dimensional sketch of a shrub or tree, with its stable trunk and more ephemeral branches. This branched design is a reflection of local resources—or their absence. Every twist in the stem of a vine, for example, is a record of its responses to changes in light and support brought about by the comings and goings of the plants around it. The location and architecture of ant trails offer a similar record.
The parallels between ant highway systems and plant architecture are easiest to detect in clonal plants such as ivy that spread over surfaces with the intelligence of an ant superorganism.34 Sometimes ivy uses what is called a guerilla strategy, developing long, unbranched stems that carry it quickly (for a plant) through sectors poor in resources—such as shady spots—with a minimal investment of tissue and little searching around. Entering a sunny patch, the ivy shifts to a phalanx strategy, combing the ground by growing more branches and short stems. It can even sense nearby plants (by shifts in the wavelength of light reflected from their foliage), and so can grow away from rivals.
Many kinds of ants similarly orient their trails to the location of food while circumnavigating the competition, as we saw with marauder ants, which construct long direct trails to distant productive regions and short branching ones within food patches.35 But unlike marauder ant colonies, which employ scoutless raids that are uncertain of what lies ahead, leafcutter colonies send lone foragers to gauge conditions over several feet and lay recruitment trails to choice vegetation. Still, because the distances they search are so short, the colony as a whole—the superorganism—is effectively nearsighted and makes choices based on distorted information. A tree next to a trunk trail may be discovered no matter how far it is from the nest, whereas a sought-after specimen away from a trail will go unnoticed, no matter how close it is to home. Even if a forager were to go far enough to detect such a plant, she’d probably be unsuccessful in mounting a return expedition. As a result, leafcutters haul foliage from long distances for what may appear to us to be no good reason.
Marauder ants and certain seed-harvesting ants have trail systems that shift every few weeks, apparently to track the location of food, but leafcutters are obstinate about retaining old trunk trails and seldom start new ones. Some ground-cover plants show a similar static pattern, staying in place like sit-and-wait predators to absorb nutrients as they become available, before their more mobile plant competitors, creeping over the ground, can show up.36 Many leafcutter trunk trails may function this way, their exact location a historical artifact of their being laid out in early life while the maturing superorganism was in an adventurous state of mind, with its workers exploring farther and in novel directions. Thereafter the routes are maintained through a kind of inertia that locks a colony into certain sectors within reach of a nest, where leaf flushes can be quickly harvested as they become available. Over time, heavy use and trail maintenance crews make the routes wide, smooth, and deep, enhancing their durability. Even if such a well-made trail becomes inactive, it’s likely to remain visible for months, if not years. This physical persistence pays off by enabling a colony to revisit sites that have shown a high productivity in the long term. The greater experience of older ants has nothing to do with these choices, because the trails last far longer than the life span of the individuals. Traces of a trail are a kind of long-term memory at the superorganism level.
To keep their trunk trails operating at capacity, leafcutter ants lay waste to anything that gets in their way. Camping in the Kaw Mountains, I was awakened by a rivulet of water next to my face. I had pitched my tent on a nocturnal leafcutter route. After nightfall, the workers had cut crescents from the floor, opening gaps in the waterproof material to allow their traffic to continue. Rain had come, and I turned on my flashlight to find my sleeping bag flecked with nylon discs and my belongings thoroughly soaked.
System maintenance, including trail-clearing operations, is as expansive as trail building. Removing a kilogram of debris from a trail takes 3,359 ant-hours of labor, equivalent to the energy content of four Snickers bars. For an ant, that’s a lot of kilojoules. Trail-clearing workers, most of which are larger than leaf carriers but smaller than soldiers, are present in sufficient numbers that obstacles such as litter—or tents—are quickly removed. They haul off small objects and gnaw larger ones while smoothing and widening the trail’s surface, until any traffic problem is alleviated. Minor workers have a separate role in trail maintenance: they loop back and forth under the feet of the larger, leaf-bearing ants to reinforce the trail as its pheromone markers dissipate.37 This is an especially important task when a section of trail is damaged by a falling branch or a passing animal or washed away in a storm; until the chemical signals are reinstated, commerce halts.
When foragers depart from the trails, they prefer to search on plant limbs, including fallen branches, rather than on the ground surface—which makes sense, given that the leaves the ants seek grow on twigs. Incorporated into trails, roots and branches serve the ants well: clean and smooth, they are maintenance free and suited for speed. By following them rather than moving along the forest floor, the multitudes of ants reduce their transit time in total by many thousands of hours over the course of a day.38
Some leafcutter trails are so well etched in the earth that I have gotten lost while hiking in the South American tropics when I mistook an abandoned ant roadway for a tidy human path. A well-built trail increases ant walking speed four- to tenfold.39 With scores of ants waving leaf banners the breadth of a thumbnail, their caravans can seem Olympian in pace and scope. “If we magnify the scene to human scale,” writes Edward O. Wilson, “so that an ant’s quarter-inch length grows to six feet, the forager runs along the trail for a distance of about ten miles. . . . picks up a burden of 750 pounds and speeds back toward the nest at 15 miles an hour—hence, four minute miles.”40
Yet with so few superhighways and so many ants on them, congestion still can be a problem. Unlike outbound marauder and army ants, which avoid those returning home with bulky prey by taking to the trail’s edges, forming one inbound and two outbound lanes, outbound leafcutters simply slip to the left or right of the homebound leaf carriers with their slim loads. Though full-stop head-on collisions are rare, each slight run-in jogs a leaf carrier off her path. Paradoxically, the best traffic flow occurs when outbound and returning workers pass one another in equal numbers, maximizing this interference, which spreads all the ants apart across the trail, forming no lanes at all.41 If this scattering doesn’t occur, the carriers end up too close together and the leaves bump together, impeding each ant’s progress. The small but frequent diversions therefore result in the fastest foliage retrieval overall, even though the ants are slowed individually.42
The two that watched the garden . . . did not notice the ants who were robbing them . . . climbing the trees to cut the flowers, and gathering them from the ground at the foot of the trees. . . . Thus the ants carried, between their teeth, the flowers which they took down . . . [and] quickly they filled the four gourds with flowers.43
This passage from the Mayan creation myth Popol Vuh describes leafcutters stealing flowers from under the noses of two guards to aid the “hero twins,” Hunahpú and Xbalanqué. It is a description of caching, in which workers deposit fragments where others have been left, a kind of positive feedback that leads to consolidated piles like those in the flower-filled gourds.44 Sometimes leafcutters leave their pile until the next day, though there’s a danger a competitor will steal it in the interim. The delay may reflect a preference for wilted foliage, which loses its chemical defenses, is lighter, and, like aged restaurant beef, is easier to chew.45 Caching is also efficient because the fragments are more likely to be transported onward when they are part of a clump than when they are abandoned in isolation.46
Caches of up to a thousand pieces accumulate when ants cut foliage faster than it can be processed or when the carriers’ progress has snarled. Workers might try for a time to enter a cramped nest entrance with their leaves, then give up and add their fragments to a stash outside. Or a worker on a subordinate traffic artery may drop her leaf when she reaches the main trunk trail, perhaps because, like a nest entrance, this juncture is a bottleneck, leading to traffic pileups at trail intersections. It’s also a sensible place for a cache because the ant carrying the leaf is likely to be familiar with the neighborhood where she found it and, once she’s deposited her burden, can return to cut again. If she continued to the nest, it would be hours before she made it back to that tree. By then, her knowledge of the local foraging situation would be long out of date.47
While leafcutters lack group transport teams, caching is an example of a different method of coordinating a workforce, one they excel at, called task partitioning: the subdivision of a job, such as the carrying and processing of leaves, into sequential stages. Task partitioning isn’t always effective. When I renewed my driver’s license at the DMV, I spent the first half hour in a line to get a number to wait in another line. But task partitioning makes sense if an overall savings in time or effort is the result. Among many kinds of ants, for example, it’s common for workers to take burdens directly from clumsy carriers. In leafcutters, handoffs from worker to worker often lead to a better match between leaf size and worker size. (I suspect this is because it takes a worker larger and stronger than a leaf’s carrier to get the clumsy ant to release her grasp.) The result of these transfers is speedier delivery.48 Some corporations have become similarly proficient at this kind of task partitioning, avoiding logjams by setting rules that mandate that employees who move faster in one step of a complex procedure take over from colleagues slower at that step.49 By contrast, an ant taking a leaf from a cache will likely move slower than the one who dropped it there because it is not easy for her to select a burden appropriate for her size from the pile.50 But despite this seeming inferiority to handoffs, caches are still common. The difference in local and large-scale efficiency between direct handoffs and transfers at caches reflects the traffic problems these techniques solve: a handoff is an immediate response of one worker to the difficulties of another (perhaps after she picked up a too-big leaf at a cache), whereas most caches are stopgap solutions to wholesale gridlock in the processing line.
Caches aren’t the only way ant colonies trade individual effort for a society-wide increase in efficiency. I once witnessed a remarkable sight deep in a rainforest near Manaus, Brazil: a rain of confetti spinning through the air beneath a tree. Peering overhead, I couldn’t make out where it was coming from, but I did see a column of leafcutters climbing the trunk. They were apparently delivering their harvest to workers on the ground via airmail. Many of the pieces were larger than an ant could carry and so plummeted straight down, minimizing the loss that would have occurred if the pieces had been small and light and liable to drift over a wide region. As it was, the contingent on the ground located only about half the cuttings, slicing them up further for transport. A 50 percent yield may be acceptable, given that the ants saved themselves the trouble of hauling the foliage down from the treetops.51
In his classic 1874 book The Naturalist in Nicaragua, the British geologist and natural historian Thomas Belt described another instance of leafcutters using gravity to save time, in this case when transporting bits of fungus garden during a migration:
I found them busily employed bringing up the ant-food from the old burrows, and carrying it to a new one a few yards distant; and here I first noticed a wonderful instance of their reasoning powers. Between the old burrows and the new one was a steep slope. Instead of descending this with their burdens, they cast them down on the top of the slope, whence they rolled down to the bottom, where another relay of labourers picked them up and carried them to the new burrow. It was amusing to watch the ants hurrying out with bundles of food, dropping them over the slope, and rushing back immediately for more.52
HUNTER AND PREY
The fact that leafcutters live on foliage and fungus doesn’t mean they aren’t as picky about their meals as meat-eating ants. Foragers may largely keep the interests of the fungus in mind—in a sense they are shopping for someone else—but because the adult workers are sustained largely by sap, some of the plants they harvest could reflect personal taste rather than the needs of the gardens. Still, the ants don’t drink the sap while they cut and carry leaves; that typically happens in the nest, where the small ants lick the fragments and regurgitate the liquid to their larger sisters.
While the colony as a whole consumes varied foliage, individual ants become specialists on certain plants growing at sites they get to know intimately.53 Workers are prompt at recruiting assistance to a plant species they know well; conversely, they recruit to an unfamiliar plant only after they assess its quality.54 In this approach they resemble a bumblebee, which, after sampling a variety of flowers, comes to specialize, or major, in a single plant species.55 It’s unclear whether majoring makes a leafcutter in any way better at her job. In any case, tender leaves come and go, and, like bumblebees (and many college students), a leafcutter worker has to change her major now and then.
There are several aspects to leaf desirability—for the ant, the fungus, and the colony. Leafcutters prefer vegetation that is easy to slice. Also, they gravitate toward foliage in direct sunlight, which is the most nutrient rich. Red leaf flushes indicate chemicals toxic to fungi,56 and leafcutters avoid them in favor of older leaves or, ideally, soft, defenseless young leaves with less of the cellulose their fungi can’t assimilate.57 Such foliage is particularly abundant in pioneer trees, which are species that spring up in early-successional habitats—relatively open places where the mature trees of heavily shaded, old growth forest have been felled by storms or old age. Where there are many pioneers, leafcutters have the luxury of selecting the few most desirable plants, whereas in older forests, colonies are forced to constantly sample from dozens of less-choice trees.58
Human land-clearing practices keep vegetation in an early-successional stage, which is why cultivated land is the leafcutters’ favorite grocery store. Many human cultivars are of Old World origin and have no native defenses against leafcutters, or they have had the toxins bred out of them for human consumption, turning them into perfect fungus garden fodder and allowing the ants to strip them bare.59 For these reasons, leafcutter populations have thrived along with human populations to a degree that can be as crippling as a biblical plague of locusts, resulting in hundreds of billions of dollars in damage annually.60
It’s curious that plants don’t do a better job of fighting leafcutter incursions. The munching of caterpillars or beetles can induce plants to produce chemical deterrents that make their tissues unpalatable or even deadly, in much the way our bodies fight a viral infection by producing antibodies.61 Plants damaged by leafcutters don’t escalate their defenses in this way, an unexplained oversight that allows colonies to harvest from the same tree again and again over the years.62 But despite the inadequacies of plant deterrents, the ants seldom completely denude full-sized native trees. Some researchers have suggested that leafcutters are “prudent pruners,” taking only a portion of each plant so it can recover for future exploitation.63 Still, it’s doubtful that ants are more sensible than humans when it comes to resources. They likely extract foliage as fast as they can, at times removing enough leaves—20 percent or more—to adversely affect a tree’s survival and reproduction.64 Yet they depart after taking the best leaves. In fact, like a kid who’s eaten enough chocolate to make herself sick, the ants can tire of certain plants, shunning a once-choice species for weeks.65 Foliage-cutting workers particularly avoid a plant species when their small comrades on the gardens detect signs of fungal ill health, suggesting that the fungus may be informing the ants of its needs.66
Foliage isn’t the only thing leafcutters take from trees. Fruit, seeds, and flowers make up the bulk of their collections during tropical dry seasons, when fresh leaves are scarce.67 Rich in calories and containing few noxious chemicals, these plant parts, which workers remove directly from plants or snatch after they fall in near-mint condition, can be more sought after than foliage. The ants also collect the pulp of fruit discarded by birds or mammals and extract seeds from animal droppings, adding any attached pulp to the fungus gardens after discarding the seeds in their garbage heaps.68
With their sweet tooth for fruit and sap, it’s a surprise that Atta workers have never been seen drinking from the sugary nectaries on plants that other ants visit so readily.69 Nectaries encourage predatory ants to protect foliage and flower rather than cutting them up, as leafcutters do; but how do the same nectaries keep leafcutters away? Perhaps their fluids contain fungicides that discourage leafcutters. No one has investigated this possibility.
Leafcutters are concerned with more than hunting down plant parts—they must guard against becoming prey themselves. Once in the early 1990s, I squatted for three days straight in the narrow space between strangler fig roots near one of the Mayan ruins at Copán where priests had once performed ritual beheadings. I was looking for a phorid fly, and I knew I had found one when a leafcutter worker threw herself back to make a quick jab at a tiny speck that appeared suddenly over her head. A phorid floats around leafcutter trails like a dust mote until it swoops down on a worker’s head, inserting an egg through the ant’s neck or mouth. Some flies even find the carried leaf fragment a convenient site to cling to while they insert an egg.70 The hatched maggot then consumes brain and muscle until finally the ant’s head falls off—hence one common name for the phorid: the decapitating fly. The worker I watched warded off her pursuer, but if she had been carrying a leaf—making her unable to move fast or defend herself—she would have been an ideal target.71
Because decapitating flies need to see the ants in order to aim for their heads, some leafcutters forage only at night, and only the workers too small to be parasitized venture out during the day. This strategy compromises productivity, however, because the smaller ants are less effective at cutting and carrying most kinds of foliage than the bigger ants that emerge after dark.72
It appears that the best strategy for dealing with these pests, like most leafcutter strategies, involves a specialized labor force. As workers cut foliage, they stridulate, producing a vibration that travels through the ant’s body to her mandibles. This causes the leaf to stiffen, but unlike with an electric carving knife, this doesn’t improve the speed and efficiency of the cut; rather, it sets the leaf itself vibrating. The better the leaf and the hungrier the cutter, the more often she chirps, which suggests the vibrations communicate the location of a good leaf to nearby workers, encouraging them to follow recruitment trails in her direction.73 Stridulation is a modulatory signal, meaning chirp intensity motivates workers to respond, much as we take cues from how feverishly a dog wags its tail.74 The vibrations become especially urgent during the worker’s final moments of sawing and her initial maneuvers to carry the leaf, at which point small workers react to her signals by climbing on the fragment as she carries it away.75 The function of these feisty hitchhikers has been the subject of much research, but primarily they serve as shields to keep the leaf carriers from losing their heads.76 The shotgun riders thrust their legs or snap their mandibles when a fly comes close. Too small for decapitating flies to target, the bodyguards become more numerous whenever the flies are abundant.
Even when a leaf carrier lacks a protective force, she can summon one quickly by stridulating.77 That’s because the small workers that ride on leaves are the same ones that reinforce trails with pheromones, during which time they patrol for threats and respond to any sign of trouble. They are also the workers that, in preparing the leaf fragments inside the nest, lick them to ingest a meal of oozing sap and, more important, to scrub off any contaminating microorganisms. It makes sense to get started on this essential task before the foliage reaches the delicate gardens—and in wide-open spaces rather than the cramped quarters of the nest.
Workers of the leafcutter ant Atta cephalotes toting leaf fragments on Barro Colorado Island, Panama.
The flies, and the occasional raid by predatory army ants, are among the biggest problems the leafcutters face. Gardening frees them from competition over food with other kinds of ants, though leafcutter colonies have been known to fight with each other along contact zones that shift back and forth like the battle lines at Gettysburg.78 Otherwise their chief competitors are solitary plant-feeding insects, few species of which have anywhere near the ant aptitude for search and seizure. Little has been written about whether leafcutters ignore, scare off, or kill caterpillars, bugs, and beetles, but in all likelihood these leaf eaters are inconsequential to them.
THE EMBRYONIC EMPIRE
Perhaps the biggest challenge any colony faces is getting started. The process is much the same for all ants. Leafcutters add a wrinkle to the story with their fungus, which is an essential part from the beginning. New queens, larger than the workers of any ant species and each the size of an unshelled peanut, tuck a wad of it into their mouths when they leave their birth nest. To pursue the superorganism metaphor, they are like so many eggs cast out by a fish, and are similarly fertilized as they disperse. (As in all ant species, they often have several midair consorts, who die immediately. Most queens don’t last long, either, but soon become food for animals and humans. Eating fried queens is like eating crunchy nuts.) Afterward, each queen will search for a place to rear her new colony. Once she digs the first chamber of her new nest, which she will probably never leave, she spits out the fungus and defecates to fertilize it. This moment must be as delicate as blowing on sparks to start a fire, for if her fungus dies, she will, too.79
With luck, in a couple of days she will have a small, robust garden. Meanwhile, she lays two kinds of eggs: small ones that develop into larvae, and large, infertile ones that serve as food for her developing brood. (Later the queen will eat similar eggs, laid by a clustering retinue of young mid-sized workers born with rudimentary ovaries that shrivel away as they get older.)80 From now on the queen serves as the ovaries for the whole—the superorganism she has created, which has, like a fish or a person, a life cycle of its own. Her workers take over the other bodily functions, equivalent to the animal organs that scientists call somatic: muscle and bone, for example. Like these parts of a body, the workers cannot reproduce, but they provide a safe environment in which the reproductive parts can create the next generation.
In its early stages, a colony undergoes a kind of embryonic development at a superorganism level.81 By the time the queen’s initial brood is old enough to leave the nest, the young society seems as precocious as a calf, with its wobbly ability to stand at birth. The queen’s first twenty to sixty workers encompass a minimum range of middling worker sizes needed to form a simplified version of the processing line of mature colonies, just enough to tend the garden and the young and to cut and process leaves, and thereby get the colony going. In the months that follow, it will grow in complexity as smaller and larger ants appear.82
With luck, the queen found a nesting spot in an open habitat. Even the recent death of one tree within a forest may provide a good-enough start. A tall tree rends a hole in the canopy when it collapses. This treefall gap lets light into the understory, allowing herbs and pioneer trees, which do poorly in deep forest shade, to move in. Even a juvenile colony’s modest labor force can process saplings of these plants, whose soft foliage is easily cut by small, less powerful workers.83 During its first, exploratory years, a colony creates temporary trails straight to such small plants within a short distance of the nest.
A growing colony will add larger road-building and leaf-cutting workers as it expands its reach to the canopy. This is to be expected: work tends to be divvied up in large societies (and in the bodies of large organisms) as a result of the differing functions or duties that must be performed in the larger, more variable area they occupy; the first complex human cultures, for example, arose where populations were dense enough for trading to become practical, and local groups could develop and maintain particular skills involving goods, such as flints for tools, specific to where they lived.84
When the larger leaf-cutting workers begin to appear, they first test their mettle on tougher tree foliage that has fallen to the ground. Finally the colony begins to build its first durable trunk trails, which never lead directly to single bonanza plants like the colony’s first feeble trails did, and the workers start to climb tall trees. It takes a couple years to add soldiers to the mix. By this time workers have begun harvesting in the high canopy and carving out giant midden chambers well before they are needed—which will be long after the crews that built them are dead, because the workers live only a few weeks.85 By its fifth year the superorganism has reached sexual maturity, producing males and queens that depart on mating flights. The superorganism can survive twenty-five years, it appears: that’s the record life span for a leafcutter queen, and there are no backups.