A Blog by Ed Yong

The Architecture of Living Buildings

The beautiful photos on this post are by Tim Nowack. Check out his site.

Fire ants are self-architects: they construct buildings using their own bodies. Put a chasm in their path, and they’ll unite into a living bridge. Tip them out of a beaker, and they’ll pour forth like a living, many-legged waterfall, with ants at the top supporting the weight of those below. And if you put them in water, as often happens in the Argentinian floodplains that they come from, they’ll merge into rafts.

The rafts are incredible. They can assemble in minutes, and stay afloat for months. By trapping air pockets and relying on their own waxy waterproof coats, the ants can float en masse, even though each of them is individually denser than water. They’re so good at floating that it’s hard to physically push them down with a stick—try it, and the raft will dent the water but won’t go under.

Can't keep a good ant raft down. Credit: Tim Nowack
Can’t keep a good ant raft down. Credit: Tim Nowack

David Hu’s team at the Georgia Institute of Technology have been studying the ants for years by filming them, dumping them in water, and measuring their physical properties. But one thing was always missing: they had never managed to actually peer inside the ants’ self-made structures, to see how they are built. “Imagine you have a thousand ants,” says Hu. “You’ll only see the ones on the surface. You can peer into the second layer but everything on the inside is inaccessible.”

The team, fronted by student Paul Foster, has now finally unveiled this hidden world. First, they swirled groups of 110 ants around in a tube to make them gather into their living balls. Then, they poured liquid nitrogen over the insects, killing them instantly and freezing them in place. They also infused the balls with superglue vapours so they wouldn’t lose shape as they thawed.

Finally, they put the ants in a CT scanner, and visualised the position of every leg and antenna, every bite and touch. “It’s like looking at a building, and seeing the scaffolding and the individual nails,” says Hu.

The scans showed that the ants are extraordinarily good at finding each other. Even though the team forced them to form into balls very quickly, while being swirled in a beaker, almost all of them attached all of their legs to a neighbour. The team studied four groups of 110 ants, giving a total of 2,640 legs. Of these, 2624 (99 percent) were stuck to another ant.

“It’s like you’re in a mosh pit, someone says, ‘Go!’ and  you have to put all your hands and legs on your neighbours,” says Hu. “The ants do that with six legs, and there are no freeloaders. There are no ants that aren’t sticking to anyone.”

On average, each ant makes 14 connections with its neighbours. They do it not with their claws, but with sticky pads at the ends of their feet; dry the pads out with talcum powder, and the ants can’t form balls, rafts, or bridges.

An ant connection. Note the adhesive pad between the legs. Credit: Foster et al, 2014. J Exp Biol.
An ant connection. Note the adhesive pad between the legs. Credit: Foster et al, 2014. J Exp Biol.

They don’t just stick their pads to the nearest thing they can find; they typically attach to their neighbours legs and feet, rather than their bodies. These connections allow the ants to change the shape of their structures by bending or stretching their legs. That explains why the structures are so elastic, and why they can absorb incoming forces more effectively.

The foot-to-foot connections also suggest that the ants actively control the nature of their balls. The team found other such clues. For example, a ball of living ants is less densely packed than a ball of dead ones, implying that they are actively pushing their neighbours away. This presumably helps to create the air pockets that keep the rafts afloat.

The living ants are also arranged differently than the dead ones. In a dead ball, the ants take on a more parallel alignment, in much the same way that rice grains or staples would if you put them in a jar and shook them around. This minimises the air gaps between them and allows them to stack more efficiently. But in a living ball, the ants are more perpendicular than parallel. This might help to space them apart and keep them afloat. It could also make the balls stronger. “If the ants were all parallel, you’d get fracture planes, and could easily break the ball in half,” says Hu. “In this pattern, you don’t have weak spots.”

A CT-scan of an ant ball. Legs have been digitally removed to make the individuals easier to see. Credit: Foster et al, 2014. J Exp Biol.
A CT-scan of an ant ball. Legs have been digitally removed to make the individuals easier to see. Credit: Foster et al, 2014. J Exp Biol.

None of this requires intelligence; there are many examples of complex animal behaviours arising from incredibly simple rules. Still, the team’s discoveries suggest that the ants are behaving in a more complicated way than anyone suspected. They’re not just randomly grabbing their neighbours. Instead, they’re arranging themselves in very specific ways. “They’re literally building a new type of material with special properties, because of the way they connect up,” says Hu. “It requires some dexterity.”

The 110-ant balls that Hu studies are also very simple. Natural rafts can contain thousands of fire ants. A bivouac (stationary shelter) of ants can control its own their own temperature and humidity by opening pores in its living walls. These structures are so big that they won’t fit inside the scanner that Hu used, so he now wants to freeze them, take them apart, and study them piece by piece. If a simple ball is already more complicated than anyone imagined, what will a tower, bridge or raft look like?

And while you’re waiting for the answers, why not enjoy a comforting cup of teOH HELL NO.

Fire ants behaving like a fluid. Credit: Tim Nowack
Fire ants behaving like a fluid. Credit: Tim Nowack

Reference: Foster, Mlot, Lin & Hu. 2014. Fire ants actively control spacing and orientation within self-assemblages. J Exp Biol. http://dx.doi.org/10.1242/jeb.093021

PS: Fire ants are named for their painful burning sting, and they are obviously very good at surmounting obstacles. “It takes some practice to work with them,” says Hu. Gloves are essential, as are Teflon-coated containers that stop the ants from climbing out. Despite these precautions, bites are inevitable. Since the team often whirls the ants into balls, individuals sometimes go flying out. “You sometimes get an ant bite many hours after the experiment because there’s an ant crawling on your clothes,” says Hu.

More on ant behaviour:

5 thoughts on “The Architecture of Living Buildings

  1. Couldn’t find the paper link in the article. I assume it’s J Exp Biol 217, 2089-2100: http://jeb.biologists.org/content/217/12/2089.abstract (paywalled, sadly).

    Another interesting bit from the abstract: “a large distribution of ant sizes permits small ants to fit between the legs of larger ants, a phenomenon that increases the number of average connections per ant.”

  2. Nice work, Ed, and lovely photos. I dealt with fire ants far too many times in Texas, and their persistence and ubiquity is incredible. One place I lived in, I you had to put any clothes you’d worn, even for a few minutes, in a (dry) claw-foot bathtub, for they are drawn to any tiny bit of sweat in the clothing and the tub was the only place they could not reach. It was constant war.

    Also, there’s a John MacDonald novel where the bad guy, tied up, of necessity, for an afternoon outside in Florida, is, alas, found dead by the hero, Travis McGee, when McGee goes back to take him to justice properly; fire ants had got him. This was a particularly bad bad guy. Made for a happy ending.

    Ed, it was wrong to run the photo of the teapot. Just wrong.

  3. Ah yes, once again, “…the team’s discoveries suggest that the ants are behaving in a more complicated way than anyone suspected.” Wonderful world. Love it!

  4. Hello Ed,
    I’d like to first state that I love your writing and am so excited for your book. However, in this post there may have been an error in the mathwork. The original research article reported that 26 legs were left unattached. I believe the number of legs attached should then be 2614, instead of 2624, from the total of 2640. I only noticed this because I am writing a science perspective for a summer science program for highschool students and I became intrigued in this subject from your article. So I read the original research article and found the numbers.
    Thank you for your hard work!(:

Leave a Reply

Your email address will not be published. Required fields are marked *