Army ant (Eciton hamatum) bridge building

In the early 1800s, Western scientific studies of wild animals involved a mix of people who (a) traveled to far-flung countries and collected/killed creatures while noting their behaviour, and (b) sat around in comfortable chairs back in Europe or the US, sorting and describing and speculating on all the field reports and dead critters.

These men—they were mostly men, thanks to rampant sexism—gathered regularly in learned societies to try and outdo each other with their findings. On 28 August 1849, medical doctor Thomas Savage stood before one such group, the Academy of Natural Sciences of Philadelphia, and read his latest paper on West African red driver ants (then Anomma rubella, now a species of Dorylus).

In it, Savage paints the ants as a cooperative and industrious species, clearing paths as the colony moves locations from day to day:

One is seen dragging along a straw or stick many times his own length and size; another grasping, rolling, then pushing along a stone far exceeding his own weight and bulk, and when his own power is not sufficient, calling in the aid of others, each knowing that a work is to be done, none idle, and every one doing promptly his part.

The effect is something like a well-drilled military unit, a point made explicitly by Dr Savage in describing various ants as either soldiers or labourers. Here you can see the soldiers scattered around the central migration path, each with their large jaws raised and ready to deter anything that strays too close:

Something else caught Savage’s eye: their cooperative spirit extended to ants using their own bodies as construction material. An 1863 summary of Savage’s work by the President of the Entomological Society in London, Frederick Smith, includes that behaviour amid a description of the ant’s unmistakable ferocity:

They kill and carry off their prey, tearing it into pieces. Their hold is very tenacious; they never leave go except when their head is torn from their body. If a small stream intercept their march, they compass it by throwing across a bridge of their own bodies, over which the whole column passes. The dread of them rests on every living thing.

Leaving aside the dread for now, it’s that bit about bridge-building that forms the basis of today’s post. The question to keep in mind is this: if an ant uses its nest-mates as a form of rope or chain-link bridge, is that tool use? Or consider it from the human angle: can our friends and family ever be thought of simply as physical objects that we use to attain some goal? Is this Dorylus soldier more than a tenacious purveyor of dread?

To investigate, we’ll need to spend time with the African driver ant’s similarly-behaved cousin, the South and Central American army ant (Eciton hamatum). And we’ll need to understand exactly what’s happening as ants form and break apart those bridges. But we’ll return to the driver ants at the end, when we—and they—come up against a much bigger and more accomplished tool-user.

A bridge too far

This is what it looks like when scientists test army ant bridge-building, by enticing the trail to cross artificial gaps:

Bridges aren’t the only thing Eciton ants build, though. As Harvard University biologist Helen McCreery reports:

They routinely form entire nest structures out of themselves (called “bivouacs”), and they dynamically and plastically control the temperature within. They also form a variety of structures to protect and ease the extremely heavy traffic flow of their foraging and migration trails. These include bridges, ladders, ramps, scaffolds, pot-hole plugs, and defensive walls, which they build over rough terrain.

However, bridges offer some of the clearest opportunities for unpicking the processes underlying ant construction. That video above shows some of McCreery’s work on the topic, testing what happens as these army ants encounter gaps in their foraging or migratory trails.

From earlier studies, we can say that none of the individual ants involved ‘intends’ to form a bridge, instead they respond to cues in their local environment, including the actions of their fellow travellers. The ants are sensitive to disruptions to the pheromone trails that guide their movements, and to the pressure of other ants walking on and around them. Experiments have also shown that there is something of a trade-off between committing too many ants to a bridge (meaning fewer ants carrying things) versus going the long way around a gap, and bridges are formed and reformed countless times as the convoy moves from one place to another. Unlike for human bridges, these ant works are very much a dynamic phenomenon.

Even after an ant bridge of the right size and strength is established, gaps can open up between the leaves or other bits of vegetation that anchor each end, as the wind and the ants themselves shuffle things around. McCreery and her colleagues wanted to know what was going through the (hive) minds of these ants under those changing and unpredictable conditions. And from the perspective of our initial question, is there anything tool-use-like involved?

Should I stay or should I go?

First, the scientists needed to hijack an army ant trail. In a passage from their 2022 report on the topic, McCreery and her co-authors describe how they enticed the swarming ants to cross an artificial platform in the forest of Panama. After placing the retractable plastic platform over a trail:

We then moved the ants’ foraging trail onto our apparatus by constructing ramps up onto and down from the platforms out of leaves and twigs that ants were already walking on (which had trail pheromone on them). We also completely removed the part of the foraging trail underneath the platforms by disturbing the leaf litter, to prevent ants from bypassing our apparatus. We physically moved ants from underneath the apparatus to the tops of the platforms, while other ants reached the platforms using the ramps we constructed.

Not mentioned: the ant bites and chaos that accompanied this process. We’ll have to imagine them, along with the dread that presumably overcame the researchers on seeing these tiny killing machines up close.

Once cameras were set to film the platform from the top and side, this is the result:

Ants walk up the leaf/twig onramps, hurry across the two white platforms, and head back down the other end. One of the platforms—on the right in the image above—has the retractable section that allowed McCreery to grow or shrink the gap the ants needed to cross, a millimetre at a time.

The simplest outcome would be that extra ants join the bridge if the gap suddenly grows, and leave as soon as the gap closes again, roughly at the same rate regardless of the how the gap changes. It would be a relatively straightforward matter of sticking around just as long as needed, no more. But that’s not what the team found.

By measuring both the volume of ants involved, and how taut the bridge was, they discovered that joining and leaving were governed by two separate processes. If you (an army ant) comes across a developing bridge that has a heavy traffic load of other ants, you should be prepared to lock arms and become part of the structure. This is because the ants invest more in those parts of the trail that are functioning well, which a high traffic level indicates. The fact that a bridge is already performing well and holding that traffic doesn’t mean you can ignore it, though, because it may still have fewer ants than it needs to be stable. Instead you should shore it up, minimising any possible ant-deficit. You should definitely join near the bridge end if the gap is increasing (whether because of shifting leaves or machiavellian scientists). Then, as long as the traffic above you stays strong, just hang in there.

So, what prompts an ant to leave as the gap contracts again? Here, the number of ants wandering over head is less important. What many of the ants will experience is slackness, as the previously sufficient number of bridge-ants becomes too many, and the excess causes the bridge to sag in the middle. Not all ants will feel that release of tension—especially those anchoring the bridge at each end—but those that do are more ready to give up being just another brick in the wall. As more ants leave, the bridge will automatically tighten up and regain an equilibrium where the number of ants joining and leaving match up, and the bridge is ready for its next disturbance. (Fun fact: the stable equilibrium state for Eciton was 0.51 ants per mm of gap, or roughly 10 ants for a 20mm gap. If you’re in the ant-bridge construction business, now you know how to budget for your next job.)

McCreery and her team summarise the process in this flowchart. You can think of it as a decision-making process or algorithm that the ants follow. And while it can seem fairly simple, it actually involves quite a complex blend of physical sensations and internal thresholds that need to be met for any given ant to switch from one behaviour to another (i.e., from walking to bridging and back again):

One particularly interesting outcome that McCreery and her team found was that the bridge formation shows hysteresis. Hysteresis is the term for when a cycling process goes back and forth, but not in a symmetrical way. For example, thermostats can have a built-in lag time so that they don’t continually flip between heating and cooling when around their target temperature. Instead, the thermostat ‘remembers’ whether it is in a heating or cooling phase and overshoots slightly before changing direction.

That structural memory and lag appears to be built into the changing size of Eciton army ant bridges. It means that minor, repeated changes of a millimetre or so don’t affect the bridge, since ants aren’t rushing away the second there’s a slight tension release. This nonlinear response buffers the bridge and ensures that the most catastrophic outcome—collapse of the bridge itself—is much rarer than would otherwise occur.

So, is it tool use?

Good question, what do you think? Tool use is fundamentally when an individual directly uses an object (whatever the object is made from) to change what they can do in the world. Does this apply here? I’d say that, even though there is sophisticated communication and structural coordination going on, army ant bridge building is therefore not tool use. Even though the ants are using each other for support, no one ant is picking up and positioning another ant as part of the bridging process. Collectively they make a bridge, but it’s a result of individual decisions made by each ant that joins or leaves the structure, within a super-structure hysteresis mechanism. If there was a master or constructor ant that placed its colleagues into position, or maybe even directed each one to its correct place, then perhaps we could make more of a case for those placed ants being tools. But that’s not what happens.

Here’s an Eciton hamatum soldier, not using tools. The ‘hamatum’ species name means ‘hooked’, by the way, which makes sense when you see those mandibles:

We can’t even say that the industrious labourer ants that clear away sticks and stones from the convoy’s path are tool users. They’re just moving stuff, not using stuff to make other things happen, much like a farmer shifting stones to the corner of a field before ploughing. Instead, the only tool use here is that shown by McCreery herself, in constructing the stick-leaf-platform device that she and her team built for the ants to follow.

However, this result doesn't rule out the possibility that animals, including humans, can use each other as tools. Most of the things that we do with metal, stone, fabric or wood tools—hitting, reaching, covering, and more—can in theory also be done with a living creature. For example, with enough dexterity, a New Caledonian crow could use a stick insect instead of an actual stick to probe for larvae in a candlenut tree. We could even explore whether a human riding and controlling a horse is using the horse as a tool. But I have to stress that there are also circumstances where using other animals or people as tools isn’t safe or humane—which is why you should never test the expression ‘enough room to swing a cat’.

But I want to see tool use!

Ok. Remember how I said we’d end up back in Africa with the fierce driver (Dorylus sp.) ants discussed by Thomas Savage? Those ants share the forest with chimpanzees (Pan troglodytes), many groups of which have learned that the ants are a great source of nutrition. However, the chimpanzees are also aware of the danger and pain that comes from simply sticking your hand into a driver ant swarm. So they make ‘fishing’ sticks that allow them to stay at a safe distance and dip the tool into the ant stream.

Naturally the ants react aggressively, biting onto the probe and clambering up it. The chimpanzee then just needs to either swipe all the ants together with its hand and bundle them into its mouth, or directly pull the stick through their lips and collect the vicious treats that way. We’ll have a more detailed look at this kind of chimpanzee tool use in a future post, but for now here’s a video taken from remote cameras set up by primatologist Cleve Hicks in the Bili-Gangu region of the Democratic Republic of Congo.

The video shows a chimpanzee mother approaching with a plant tool already prepared and in her mouth, before she stops and fishes for driver ants by the side of the path. In the second part of the clip, from a different camera, possibly the same female passes by with her tool, before the second chimpanzee is startled by (and then becomes quite curious about) the camera itself:

There you go: tool use with, but not by, swarming ants. And that’s where we’ll leave it today. Unless you really want to see driver ants attacking things, in which case the video below is for you.

Further viewing

I’ve saved this video for the end because it’s not for everyone. BBC Earth takes you VERY close to driver ant swarms, and then shows them overwhelming a variety of larger prey, including a scorpion. You’ll also see—again, in close-up detail—what happens when one of those soldiers gets its mandibles into a human toe. This video is the reason I learned the word myrmecophobia.

Sources: Savage, T. (1849) The Driver Ants of Western Africa. Proceedings of the Academy of Natural Sciences of Philadelphia 4:195-204. || Smith, F. (1863) Observations on Ants of Equatorial Africa. Transactions of the Entomological Society of London 3rd Series vol. 1:470-473. || McCreery, M. et al. (2022) Hysteresis stabilizes dynamic control of self-assembled army ant constructions. Nature Communications 13:1160. || Hicks, T.C. et al. (2019) Bili-Uéré: A Chimpanzee Behavioural Realm in Northern Democratic Republic of Congo. Folia Primatologica 90:3-64.

Second image credit: JB; https://freewheely.com/2014/02/across-columns-of-driver-ants/ || Third image credit: AntWeb, California Academy of Science; https://www.antweb.org/specimen.do?name=casent0172646&shot=p1&project= || Fourth and fifth image credit: McCreery et al. (2022) || Sixth image credit: AntWeb, California Academy of Science; https://www.antweb.org/description.do?species=hamatum&genus=eciton&rank=species