The science of ants' underground cities

Credit: Pixabay/CC0 public domain
Imagine an anthill. What do you see? Is it a small mound of crumbly dirt and sand poking through the lawn? A small hole in the ground. A few ants are busy. It's not very impressive, is it?

However, if you look below the surface, the simple above-ground structure gives way to subterranean complexity. Tunnels descend, branching to lead to chambers that are specialized for colony queens, nurseries for their young, and farms for fungus. These tunnels are more than just burrows. These underground cities can be as deep as 25 feet below the ground and often last for many decades.

This type of construction is a difficult one for most animals, but it's especially impressive when done by animals smaller than your fingernail.

A Caltech research team has discovered the secret to how ants create these complex, stable structures.

Jose Andrade, George W. Housner Professor in Civil and Mechanical Engineering, led the research team that studied the digging habits and discovered the mechanisms that guided them. In a paper published by Proceedings of the National Academy of Sciences, the research was described.

What do ants think (if anything)

Andrade is the Cecil and Sally Drinkward Leadership chair and Executive Officer for Mechanical and Civil Engineering. He wanted to know if ants actually "know" how tunnels are made. Or are they blindly digging.

Andrade shares that they were inspired by exhumed ant nests, where they pour plastic or metal into them. They can create tunnel systems that are incredible. "I saw one of these structures next to a human and thought, 'My goodness! What a magnificent structure.' I was intrigued to see if ants know how to dig.

Andrade states that although we didn't interview ants to find out if they knew what they were doing, we did have the hypothesis that they dig in a deliberate manner. We speculated that ants might be playing Jenga.

He meant that the team was "playing Jenga", which means that they suspected that the ants were digging in the dirt looking for loose soil grains to remove. This is similar to how a Jenga player checks for blocks that can be removed from the stack. The blocks that are unable to be removed, the ones carrying the stack's load, are part of the structure's "force chains", the collection of pieces that have been pressed together by the forces.

Andrade states that they hypothesized that the ants could sense force chains and avoid digging there. "We believed they were tapping soil grains, which could have enabled them to assess the mechanical forces they faced," Andrade said.

Ants will do whatever they want

The team had to have ants to study in order to learn more about them. Andrade is not an entomologist, but an engineer. So he sought the assistance of Joe Parker, an assistant professor of biology, biological engineering and biology whose research focuses primarily on ants and their ecological relations with other species.

Parker states that Jose and his team needed someone who can work with ants and understand the adaptive collective behavior of social insects so they could give context to what they were doing.

Parker joined the team and they began to cultivate ants. They also learned how to work together with them. Andrade said that the process took almost a year. They had to raise enough ants for work, and there was much trial and error in getting them to dig small cups of soil they could load into an Xray machine. They determined the best size cup and how many ants they should use. However, the researchers were not always able to rely on the ants' cooperation.

Andrade said that they are "sort of unpredictable." They dig wherever they like. These ants would be placed in a container and then some would begin digging immediately. They would make amazing progress. Others would take hours to dig and wouldn't bother. Some would continue digging for a while, then stop to take a break.

Once the ants started to move, researchers would then take the cups and X-ray them with a technique that creates a 3D scan all the tunnels within. Researchers could make simulations of the progress that ants made by taking several scans and letting them work a bit between each one.

Understanding ant physics

The next step was to analyze what the ants were doing while they worked. There were a few patterns that Andrade's group discovered. Andrade said that the ants were trying to be as efficient as possible. Because the cups would be used as part of their tunnels' structure, they dug their tunnels around the edges of the cups. Their tunnels were also as straight as they could.

Andrade says, "That makes sense since a straight line is a shortest path between two points." It shows how efficient the ants are at what they do, with them using the sides of the container.

The angle of repose is also what the ants used to dig their tunnels. This angle is the highest angle at which granular materials, made up of individual grains, can be piled up before they collapse. Imagine a child building a castle out of sand at the beach to illustrate the angle. The child will use dry sand to build the pile. Every scoop of sand added will slide down the sides. The pile will get taller but not wider as more sand is added. It will also become steeper. If the child uses wet, the pile will be tall enough to make walls and towers. Wet sand has higher angles of repose than dry, and each granular material has its own unique angle. Andrade claims that ants can determine how steep the angle is for whatever material they are digging in. They don't exceed it. He also believes that this makes sense.

He says, "If I am a digger and I survive, my digging method is going to align to the laws of Physics, or else my tunnels will collapse and I'm likely to die."

The team finally discovered something that could be used to help humans understand the physics behind ant tunnels.

Subtle changes in the forces chains around the tunnel are caused by ants removing soil grains. These chains are somewhat random before the ants start digging. They then rearrange themselves around outside the tunnel, much like a liner or cocoon. Two things occur as they do this: 1. The force chains reinforce the tunnel's walls and 2. The force chains reduce the pressure on the grains at the end of the tunnel, where the ants are operating, and make it safer for them to be removed.

Parker said, "It has been a mystery both in engineering and in ant ecology" how ants create these structures that last for decades. We found that ants can remove grains from the pattern we observed and reap the benefits of these circumferential force chain as they dig down.

What about the hypothesis at the core of the team? Is it possible for ants to be aware of the things they do when they dig?

What ants know, and what they don't

Andrade said that what they found was that they didn’t seem to ‘know’ what they were doing. They didn't look for soft spots in sand. They evolved to dig according the laws of physics.

Parker refers to this as a behavioral algorithm.

He says that the algorithm is not possible in a single ant. It's the emergent colony behavior that all these workers are exhibiting acting as a superorganism. We don't know how this behavioral program spreads across all the tiny brains that make up the ant colony. It is amazing.

Andrade stated that he plans to work on an artificial intelligence approach to mimic that behavior algorithm in order to simulate the way ants dig on computers. Andrade said that part of the emulation will be to determine how to scale ant Physics for human-sized tunnels.

He says that grains scale differently to other materials, such as fluids and solids. You can scale the intergranular friction coefficient to scale experiments from the grain scale (in this case, a few millimeters).

Next, what? Robotic ants capable of digging tunnels for people.

He says, "Moving granular material is extremely energy-intensive and expensive. You always need an operator to run the machines." This would be the last frontier.

The paper, "Unearthing 3D real-time ant tunneling mechanics", describes the research and appears in the August 23rd issue of Proceedings of the National Academy of Sciences.