ants have an even more efficient way to flag a route that has ceased to be profitable, signaling “don’t bother” by depositing a different pheromone at the start of the trail.)7 The chemicals required to convey a message can be minuscule. With one species of leafcutter ant, a thousandth of a gram of recruitment pheromone—a minute fraction of one droplet—would be enough to lead a column of workers around the world sixty times.8
Since traffic depends on pheromone strength, it is modulated by the ants’ overall assessment of a trail’s offerings—what we call mass communication. This technique can lead to what appear to be deliberate choices by the colony, despite the ignorance of the individual ants of such matters as the size of the food item they are visiting and the number of workers needed to harvest it. For instance, a colony will more quickly exploit a nearby food source than one farther away, simply because it takes less time for the ants to walk the shorter distance. This results in the quicker accumulation of the trail pheromone, which in turn attracts more ants to the meal.9
Among the Sullia oil palms, however, such subtleties of individual reaction and mass response were hidden to me. Instead I recorded seemingly spontaneous eruptions of ant multitudes followed by sudden mass retreats, like an arm that was extended, pulled back, and then extended somewhere else. What was going on?
I thought back to a similar eruption involving army ants that I had seen as an undergraduate studying butterflies in Costa Rica. I was awakened one morning to a rustling sound in the room of the hacienda where I was a guest. Eciton burchellii army ants were everywhere, moving in waves over the floor, flowing through cracks in the wall, falling from furniture while clinging to the backs of beetles and silverfish. I heard a plopping sound as an inch-long body landed on the carpet next to my bed: a scorpion cloaked in ants had dropped from the ceiling. The only reason no ants had swarmed my body was that each leg of the bed had been set in a dish of oil by the owner’s wife. Thank heavens—I doubt Señora Perez would have appreciated my dashing naked, draped only in ants, into her parlor. I put on my robe and ran to the ant-free hallway, then waited out the ant raid over toast and scrambled eggs.
What had those ants been doing? In a word, they were foraging.10 For all ant colonies, this search for food is carried out by multiple workers at once. But while the foragers of most ant species operate independently of each other, army ants forage together, much like a pack of wolves looking for elk.11 Unlike a wolf pack, however, army ant hunting groups do not have a circumscribed membership. Thousands may be present in a raid, but different workers come and go en route to the nest, a search strategy called group foraging or (my preference, because there is no set “group”) mass foraging.12
Many fierce predators dispatch difficult prey without searching for it in a group. In certain ant species, workers acting alone can both find and kill small vertebrates. Workers of one Brazilian ant dispatch tadpoles larger than themselves.13 But most predatory ants cannot overpower such prey without help. Most commonly, a successful forager—called a scout when a few scattered individuals are doing the reconnaissance—recruits a raiding party, often guiding it for many meters to the specific site where she discovered the prey.14
Elsewhere in the Western Ghats of India, I saw this system used by Leptogenys. The tight pack of slim, glossy ants was moving through the dry litter at the reckless speed of an Indian bus driver. I followed and watched as they entered the mud galleries of a termite colony. The ants soon emerged, each with a stack of termites in her jaws. This regimented form of group predation was a joy to observe, as long as I stayed back far enough to avoid the needlelike stingers that Leptogenys use to immobilize their prey in one-on-one combat. Later I determined that this species employs scouts. These individuals then return by themselves to the nest and recruit a few dozen nestmates who together do the potentially dangerous work of mining and transporting the unwieldy termites.
Army ants employ a completely different foraging technique. Rather than proceeding with guidance to food already found, the workers sweep ahead blindly in a mass, the absence of a single target turning the whole raiding business into a gamble.
Some army ants regularly invade homes, and in the underdeveloped world their arrival is welcome (even though they force everyone out for an hour or two), for they clear out vermin such as roaches and mice. Marauder ants perform a similar service—though they also make a nuisance of themselves by absconding with grain and other human foods, as Captain Bingham recorded.
Indeed, it was impressive to watch marauder ant mobs take on centipedes and frogs in India. From those clashes I saw that, like army ants, marauders recruit members explosively as each prey item is found, then kill and cart off the bodies together. But to understand marauder ant foraging behavior, I needed to learn how they located their prey in the first place.
After a time, Raja tired of the ant bites and stayed home to practice his guitar. By then I had been in Sullia three weeks, surviving on sticky rice splashed with a red curry so spicy that it often left me panting. This diet kept me ravenous, and to sustain my energy I purchased caramels at a roadside stand. (The shop had more ambition than inventory, with the former evidenced by its name, Friendly Mega Supermarket Store, which was crudely painted on a board.) I surreptitiously devoured the candy at night, fearful of hurting my host family’s feelings, and disposed of the paper wrappers down rodent burrows on the plantation.
One cool evening as I watched marauders rushing in the tree litter, as greedy for high-calorie food as I was, a vision of the ants as a superorganismic being crystallized in my mind. I began to think about the army ant stratagem of foraging in a “group.” Within the superorganism, what does membership in such a group entail for an ant? Is it proximity? Among humans, techniques as old as jungle drums and as new as Twitter allow people to form groups without physical closeness. Conversely, being close to others does not automatically confer membership in a group in a meaningful way. Often enough I have joined a crush of people on a city street—quite a crowd, but not much of a group.
I had seen many ant species in which nearby workers show no semblance of joint action. Is proximity even less meaningful to ants than to people? In many ways, yes. The workers of most ant species cannot detect another ant’s presence until they are virtually on top of each other. Army ants, legally blind by human standards, sense a nestmate only during fleeting moments of contact. In such times, the ants distinguish friend from foe, but what they learn is unlikely to play a role in the organization of their armies. Rather than responding directly to others, ants tend to react to information left by nestmates who may be long gone—to the webwork of social signals, such as pheromones, spread throughout the environment in an ant version of the Internet.
Think of household ants following an odor trail to a cookie left on a kitchen counter. What happens if I pluck out all but one ant? Her actions won’t change an iota as long as she can track the scent. She continues to participate in a group effort to harvest food whether the trail is thick with ants or not. Could we define an ant as being part of a group when her actions are constrained or guided by the varied signals and cues arising from the actions of her nestmates, and as solitary when she acts on her own?15
As it turns out, army ants conform to this view of a group. The workers have negligible freedom to wander far from nestmates and any fresh chemical communiqués those nestmates have left behind; the superorganism never sends out lone pieces of itself, but droves of workers operate as an almost tangible appendage that stays attached, through a continuous flow of ants, to the main body. Some scientists point to other aspects of army ant life, such as their ability to catch or retrieve prey in groups, but it is this aspect of their behavior—how they forage, and not what they do after they find food—that sets army ants apart from other ants.
My goal became to determine whether the marauder ant uses the army ant group approach to hunting. In India, I documented the movements of teeming battalions, with the workers numbering in the tens of thousands. But such details as whether the raids relied on scouts
