What I discovered in the field was that the slender minor workers form 98 to 99 percent of the population. Tiny, with heads about 0.6 millimeter wide, they are distinct. There are no intermediates between them and the other ants, which range widely and continuously in size. Within this continuum, there is a distinct peak in the numbers of ants at just over 2 millimeters’ head width, and so these I called “media workers,” and another peak at just over 3 millimeters’ head width, for the majors. A few of the majors are substantially larger, with heads 5 millimeters wide or more—the size category I informally called the giants. The queen, who ordinarily stayed in the nest, had a smaller head than a giant, but a much larger body: she could be about 2 centimeters long.
Among different kinds of ants, I learned, work is divided up in two ways. In some species the workers are similar in appearance but flexible in their job skills, temporarily taking on any tasks as they arise, but the colonies of other species can also develop workers of different sizes to do different jobs on a more permanent basis. The former method allows colonies to adjust more rapidly to changing conditions, but it has its limitations: since the workers are identical and interchangeable, duties that require a specialized skill set may be poorly executed. Polymorphism—variation in size and shape, along with physiology and brain development—is an indicator of a more permanent specialization, and is the primary determinant of division of labor in the marauder colony. Because the workers of differing size are suited to a narrower set of tasks, they expand their activities, if at all, only under stress; in some ant species, for example, soldiers who ordinarily do little except fight will help tend the brood if other workers are taken away by a meddling researcher.3
From this, it has been determined that an extremely polymorphic species like the marauder ant is likely to have predictable labor needs, because the number of members in each physical caste, or size group, changes slowly, if it can be changed at all, based on the colony’s requirements.4 In fact, the size frequency distribution reveals something about how many ants of each caste a society requires, somewhat equivalent to the distribution of people in different job descriptions in a city.5
To pursue again the earlier metaphor, a colony can be seen as a “superorganism” that functions like the body of an organism, with the number of castes and the frequency of each being analogous to the number of types of cells and tissues and the size of organs. Ant species with small colonies are like the cells in simple organisms in that they have few labor specialists, but marauder ants are intricately specialized. Add the arrangement of the workers in space and their interactions with each other to the numbers and frequencies of the various workers, and one has the “scaffolding” of the superorganism, much as a body is built upon the number, location, and interactions of cells. The parallels are all the more remarkable since both the ant workers in a colony and the cells in a body communicate largely by chemical cues (hormones being a prime example for cells), the biggest difference being that workers are mobile and accumulate dynamically when and where they are needed, while most cells are fixed in place within the body.
Essentially all the participants in the raid front are the little minors. With my photographs, I was able to disentangle the blur of action as these ants brought down a nightcrawler or grasshopper thousands of times their weight. A single minor worker has no more chance of catching such a behemoth on her own than would an equally small worker of a solitary-foraging ant species. But she shares the front with other minors that contact prey at about the same time, and they pile on like tacklers in a game of American football. With this strategy, the chances of capture improve markedly: as in Swift’s tale of Gulliver toppled by the Lilliputians, strength in numbers can’t trump size.
It makes sense for a colony to produce a lot of minor workers and concentrate them at the front. If the prey were confronted by a single media ant instead, even one weighing as much as all those smaller tacklers combined, the larger worker would be less effective at subduing the worm or grasshopper. Though individually weak, minors working together simultaneously grab their quarry at different places and angles, making it hard for a victim to move. The prey is also more likely to slip by a single big worker than by a barricade of spread-out small ones.
Countless times I’ve watched a nightcrawler inching over the ground or a grasshopper resting on its green blade, minding its own business, as a swarm moves toward it with a whisper like a snake in the grass. If it doesn’t respond by reflex, death is certain. At the touch of the first worker, the worm flips back and forth; the grasshopper makes its leap. But out of view in the vegetation, more ants are swarming in. About half the directions the flipping worm or leaping grasshopper could choose will land it deeper among the ants, while the other half will allow it to evade the ants by getting ahead of the raid. Blundering deeper is like colliding with a dragnet with a mesh of the width and strength approximated by the closeness and size of the ants; the more the worm or grasshopper struggles, the more the masses converge on it, as other ants are alerted and drawn into the fray. Soon all the little ant jaws hold their prey taut.
Avoiding the ants by moving ahead of the raid provides a temporary respite. The best hope for any creature is to dash to freedom to the left or right of the raid, and so carry itself out of the ants’ path; but the distribution of ants must be difficult for prey to determine down among all the litter and plants on the ground, so taking this course may be a matter of chance. If the prey fails to chose the right direction, the army will advance to its new location and strike again. And if it escapes once more, a swarm may try a third time, or more. Because of their width, swarm raids are most likely to repeatedly contact the flipping worm or leaping grasshopper. (A narrow column raid is different; its net is too narrow and weak, and most victims break free. The ants in column raids therefore reap mostly seeds and frail prey, though the raid may burgeon into a swarm if they find bigger spoils.)
Minor workers at the front of a marauder ant raid in Singapore being cut to pieces while subduing a termite soldier.
A major worker crushing the termite after the minors pinned it down.
Even escapees may not survive. I once saw a cricket rocket from its hiding place beneath a leaf. In a series of zigzag moves it ended up far from the raid, but a few ants still clung to it stubbornly. Their gnawing slowed it down, until at last its body convulsed. However, the ants that subdued it were now so far from their colony that they would die before ever finding their nest again.
Participants in a marauder raid seem to be forever in battle mode. They fight with a dogged precision that is chilling, and in large raids there certainly seem to be troops to spare. The minors show by far the highest casualties. The bounding cricket managed to chew a couple of the minors on its leg to a pulp before succumbing to the rest. On my way back to the raid, I saw minor workers puncturing a plump caterpillar, and one drowned in the jelly that oozed from it. Later on in that raid I saw a termite soldier with a burnished red head that dwarfed the minor workers surrounding her like a grizzly bear cornered by dogs. The termite’s black jaws were sharp as knives, and each minor that came near was sliced apart as cleanly as if by a guillotine, until a dozen ants stormed her hindquarters and brought her down.
Like a war correspondent inured to tragedy, I watched hundreds of minors being sundered and smashed in struggles with prey, the horror of the slaughter magnified through my camera lens. By never straying from the task to save themselves, they displayed breathtaking devotion to their duty. It made me wonder about the advantage of psychological numbness in combat even among sentient humans. As one author wrote of the Civil War, “Soldiers perhaps found it a relief to think of themselves not as men but as machines.”6
Such thoughts reflect how caught up I was in the drama of the moment, pressing the button of my camera each time a surprising event happened. I saw that the minor workers were able to stretch the legs of the termite soldier until she was spread-eagled (click). By this
