A new study explains how ant rafts can transform: shrink, expand, or grow into long, proboscis-like protrusions. The scientists’ findings could help develop robots that work together in “swarms” or next-generation materials in which molecules migrate to fix defects.

“The origins of this behavior are quite simple to explain. Lonely ants are not as smart as you might think, but together they make very intelligent and resilient communities,” said Frank Wernery, author of the new study and professor in the Department of Mechanical Engineering at the University of Colorado.

Fire ants (Solenopsis invicta) form giant floating clumps of squirming insects after storms in the southeastern United States to survive in raging waters.

In their study, Werneri and lead author Robert Wagner used mathematical simulations and models to try to figure out the mechanism behind these “lifeboats”. They found, for example, that the faster the ants moved on a raft, the more those rafts would expand outward, often forming long ledges.

“Basically, this behavior can happen spontaneously. It is not necessary for the ants to obey any centralized decision,” Wagner noted.

Wernery and Wagner threw thousands of fire ants into a container of water with a plastic rod in the middle, simulating a lone reed in the midst of turbulent waters.

“We left them there for 8 hours to observe the long-term evolution of these rafts. As a result, we saw that the rafts began to form a kind of growths, ”said Wagner.

Instead of maintaining the same shape over time, the structures contracted and tightened, forming dense circles of ants. In other cases, the insects fanned out like pancake batter in a frying pan.

The scientists said that the ants seem to modulate these shape changes through a process called “treadmill”. As Wagner explained, each ant raft consists of two layers. At the bottom you can find “structural” ants that cling tightly to each other and form the basis. Above them is a second layer of ants that walk freely on top of their counterparts in the colony.

Within a few hours, ants from the bottom may crawl to the top, while free-roaming ants will sink down to become part of the structural layer.

“The whole thing is like a donut-shaped treadmill,” Wagner said.

In the new study, he and Wernery wanted to find out what makes a treadmill spin.

To do this, the team created a series of models that essentially turned the ant raft into a complex game of checkers. The researchers programmed approximately 2,000 round particles, or “agents”, to replace the ants. These agents couldn’t make decisions for themselves, but they followed a simple set of rules: fake ants, for example, didn’t like to bump into their neighbors and tried not to fall into the water.

Wagner and Wernery found that their simulated ant rafts behaved very much like real ones. In particular, the team was able to fine-tune the activity of the agents in their simulations: were individual ants slow and lazy, or, on the contrary, energetic and agile? The more the ants moved around, the more likely it was that they formed long appendages that protruded from the raft – a bit like people heading for the exit in a crowded stadium.

Wagner suspects that the fire ants are using these extensions to probe their environment for logs or other areas of land.

Researchers still have a lot to learn about ant rafts: for example, what causes ants in the real world to switch from sedate to lazy? But for now, Wernery says engineers can learn a thing or two from fire ants.

“Our work on fire ants will hopefully help us understand how simple rules can be programmed, such as algorithms that determine how robots interact with others, to achieve a targeted and intelligent swarm response,” he said.