1976 | Maryland, USA

Myrmicine ant (Aphaenogaster sp.) liquid carrying

One of the themes of this blog is that to truly understand animal technologies, we must look closer. Today we’re taking that literally, as we peer into the undergrowth to find ants that drop pieces of soil, sand and leaf litter onto liquid food, as a way of collecting and transporting it back home.

Our story begins almost 50 years ago, with a couple of Fellers exploring the woodlands of College Park, Maryland, just across the border from Washington D.C.

Pick up lines

When I say Fellers, I’m referring to zoologists Joan and Gary Fellers, of the University of Maryland. As they reported in the journal Science in 1976, the Fellers were particularly interested in what local ants of the genus Aphaenogaster were up to.

This genus is common in the hardwood forests of eastern North America, as well as other similar forested environments through southern and central Europe, and across into southern Asia and Australia. Nests each have hundreds of worker ants that spread out and forage for food, which they then bring back to support both other workers and their nestmates that remain behind to tend to the queen and eggs. If you’ve been out walking in any of those places, chances are you’ve stumbled across a busy Aphaenogaster or two.

Ant tool use was known at the time of the Fellers’ report, for example in a short 1972 note on the harvester species Pogonomyrmex badius in Florida. That experimental study found that the ants would place sand grains into a spot of honey, before returning to collect the soaked grains a short while later. What the Fellers brought to the table was an admirable level of rigour, and that ability to look closer.

First, they laid out small pieces of liquid jelly bait for the woodland ants. Members of the species Aphaenogaster rudis that encountered the jelly would spend a few seconds at the bait, then wander away and return with leaf fragments. The Fellers recount:

As leaf fragments accumulated, ants from the same colony tended them, adjusting the positions of the leaves or sometimes pulling the leaves off completely and repositioning them. By individually marking ants with small spots of paint, we were able to determine that a given individual may bring several leaves and that ants tend leaves brought by other ants.

In other words, leaf-dropping ants would either head off to get more leaves, or stay at the jelly and move leaves around on the bait. These jobs were shared between colony members. Between 30 minutes and an hour later, the ants started taking the jelly-leaves directly home, travelling up to 1.5 metres away (a decent journey for the half-centimetre long ants).

In the image below, from the Fellers’ report, ants can be seen successively carrying, placing, tending to, and then removing leaves from one of the woodland jelly baits.

So far, so tool-use. There remained an open question about just what was happening inside those home nests, though. Were they actually used for food, or some other purpose? So the Fellers set up a small transparent ant farm, inhabited by related Aphaenogaster fulva ants.

They found that at least one of the jelly-covered leaf pieces was indeed taken directly to the nest chamber housing the colony queen, with her eggs and attendants (attend-ants? Sorry). Over a period of several hours multiple ants fed from the leaf, scraping the food off using their mandibles.

The quantity of transported food wasn’t minor, either. With extreme patience, the Fellers weighed dozens of leaf pieces that had already been selected by A. rudis workers (they stole them from ants who were on their way to some bait), and placed them on a natural equivalent of jelly: the squashed bodies of beetle larvae. After half an hour, each fragment gained around 1.4 mg of juicy fluid—slightly more than an ant’s own body weight.

Working even more carefully, the Fellers found that after an hour of feeding directly on jelly, the ants only gained 0.13 mg in weight. The leaf tools therefore gave them 10 times their internal carrying capacity. And it wasn’t all jelly and beetle bits either, since the ants also made good use of a squashed spider and rotten fruit pulp.

They may not have been very picky about what they ate, but it turns out that A. rudis workers can get quite picky about their tools. The Fellers noticed the ants would often pick up and then discard different pieces of material while searching for a suitable tool. By giving them a choice of mud chunks, pine needles, leaf pieces and bits of wood (all similar in size, around a half centimetre or less) in an experiment, they were able to check for ant preferences.

Mud was the top choice, used significantly more often than random chance would suggest. Leaves were actually least preferred in the experiment, which could mean that the leaf-using ants out in the woods weren’t able to find anything better. Not only are the ants tool users, but they appear to be tool selectors, able to choose which parts of the environment they want to incorporate into their dinner-tray routine.

The Fellers couldn't resolve whether tool use offered more than an efficient way to feed an ant family, but they did offer one additional idea. Aphaenogaster isn’t the only ant in the forest, and it can end up subordinate to other species that more aggressively monopolise found food. What if, the Fellers asked, tool use is a way for ants in this species to capture food before it gets overrun by the bullies? The ants can then retrieve the soaked tools at a safe time, since the domin-ant species will eat around the tools and leave them behind.

This competition idea hasn’t been rigorously tested, but it remains a possibility. It may even be that rapidly covering liquid food allows Aphaenogaster to hide it completely from other ant species.

Ant-icipating critiques

Tool use by these ants seems unequivocal: they encounter a problem, and select and manipulate a piece of the environment to extend their abilities and resolve the problem. What wasn’t clear from the early reports, though, was whether there was any flexibility to this behaviour, or even whether the tool use was incidental to an ant’s need to avoid falling into a liquid trap.

These issues weren’t top of the ethology checklist in the 1970s. In recent years, however, teams in both the US and Europe—especially at the University of Szeged in Hungary—have taken up the task.

On the question of whether ants are using debris simply to block a liquid hazard, the answer is a resounding no. In a report from December 2020, the Hungarian team worked with another related species, Aphaenogaster subterranea, to show that ants weren’t simply covering over the liquid food, then separately just happening to take those pieces back to their nest.

Building on earlier work by Gábor Lőrinczi, his student Gábor Módra and colleagues carried out an elegant series of experiments that teased apart the issue. They found that ants were far more likely to drop tools into a honey-water solution than water alone, and that honey droplets further from the nest (where they would be less of a hazard) actually saw more tool-use than those nearby. Those objects placed into water alone—fewer than 5 per water drop, compared to the 20-30 used on honey-water—were never subsequently picked up and transported.

A second experiment tested whether the ants were simply gathering up liquid-y bits of material as they found them. The scientists put a dozen or so bits of pine needle into drops, again comparing honey water to pure water. Ants added extra pieces to the honey solution, then carried them back to the nest as before. They didn’t add anything to the water, despite it remaining a potential hazard. And when the researchers tried using extremely small droplets that would not have hindered the ants, they still dropped pieces on the honey, and not the water.

What about the idea of flexible behaviour, then? Were those ants choosing mud over leaves in the Fellers report actually being choosy? Again, the evidence falls in favour of the ants.

Lőrinczi and his co-workers gave a selection of natural materials to ants in their experimental colony. You can see them here, labelled as SSG (small soil grains), LSG (large soil grains), PN (pine needles), and L (leaves). To see if the ants would accept an upgrade, they also included small pieces of human-made sponges (SP):

Over a series of experiments, the ants preferred the most easily carried tool material: small soil grains. However, they were willing to use heavier and larger items if they were placed closer to the food than the other materials. This kind of distance-effort trade off is seen in other tool-using animals, including wild capuchins that use stone tools in Brazil. We’ll check in on those monkeys another time.

Unexpectedly, the ants were also able to turn a large tool into a smaller one. Groups of sometimes over a dozen workers gathered to tear up the sponge pieces, as well as some of the leaves. Whether or not you’d be willing to call this tool manufacture, it’s certainly a step towards a more flexible approach to tool use. As it happens, I would be happy to consider this tool-making, given that the ants are reducing an existing object to make it better for the task at hand, just as a chimpanzee may trim leaves from a termite-fishing tool, or a human ancestor might chip stone flakes away from a larger core to shape the resulting chopping tool.

A personal touch

In case you’re wondering, its not just Hungarian wonder-ants that favour tool use over blocking potential hazards. At the turn of the millennium, back in Vermont in the eastern USA, A. rudis colonies passed the same kinds of tests. Both in the wild and in the lab, these ants avoided non-food liquids (petroleum jelly and water) while targeting grape jelly drops.

The Vermont ants also showed the kind of tool-making behaviour seen in Hungary. As Valerie Banschbach and her team watched:

We noted ants cutting pieces of debris from leaves or sticks immediately adjacent to the food resource, in the field, and then dropping the debris a very short distance into the food.

These scientists decided that their ants were most likely showing a kind of specialisation, where only some of the individual foragers were regular tool-users. Whether this was a learned or genetic role wasn’t clear, and the Vermont team’s experiments weren’t designed to unravel that knotty question.

However, Aphaenogaster is an international genus, and re-crossing the Atlantic we find that the notion of ant tool-use specialisation was taken up with yet another species, this time A. senilis in Spain. Again led by a University of Szeged researcher, Istvan Maak, the target was something you may never have considered: ant personality.

The scientists tested individual ants on their willingness to wander around in the open, and to explore when given new prey (a freshly killed fruit fly). Those that spent more time doing so can be considered more ‘outgoing’ than ants that preferred to stay sheltered and less conspicuous. The key finding, though, was that only a small subset of foraging ants in a colony were tool-users, and that subset were the more exploratory ants. The link isn’t fully fleshed out, but it’s possible that these more active ants are more likely to start using tools, at which point the rewards make it a self-reinforcing behaviour.

So, we end up with a specialised set of outgoing ants, actively selecting, making and using tools to collect much more food than they could consume alone, potentially set against a backdrop of inter-species ant competition for the tasty, squishy bits of the world. That’s quite a complex pageant. It’s also likely happening regularly in a forest near you, which means it’s a phenomenon you can actively explore, should you be willing to just look closer.

Sources: Fellers, J. & G. Fellers (1976) Tool Use in a Social Insect and Its Implications for Competitive Interactions. Science 192: 70-72. || Lubertazzi, D. (2012) The Biology and Natural History of Aphaenogaster rudis. Psyche 2012: 752815. || Morrill, W. (1972) Tool using behaviour of Pogonomyrmex badius (Hymenoptera: Formicidae). Florida Entomologist 55: 59-60. || Módra, G. et al. (2020) Protective behavior or ‘true’ tool use? Scrutinizing the tool use behavior of ants. Ecology and Evolution 10: 13787-13795. || Lőrinczi, G. (2014) Some notes on the tool-using behaviour of the ant, Aphaenogaster subterranea (Hymenoptera: Formicidae). TISCIA 40, 17-24. || Lőrinczi, G. et al. (2018) Which tools to use? Choice optimization in the tool-using ant, Aphaenogaster subterranea. Behavioral Ecology 29: 1444–1452. || Banschbach, V. et al. (2006) Tool use by the forest ant Aphaenogaster rudis: Ecology and task allocation. Insectes Sociaux 53: 463 – 471. || Maak, I. et al. (2020) A small number of workers with specific personality traits perform tool use in ants. eLife 9: e61298.

Main image credit: Imola Bóni, https://blog.oup.com/2018/10/ants-pick-tools-foraging-beheco/ || Second image credit: Fellers & Fellers (1976) Fig. 1 || Third image credit: https://i.imgur.com/oGy3PLE.jpg || Fourth image credit: Lőrinczi et al. (2018) Fig. 1