To Form Swarms, Midges Fly in Spring-Like Zigzags

(Inside Science) -- You're enjoying the sunset on a warm summer's day, when suddenly, a buzzing mass of midges surrounds you. The bugs collide with your face as their apparently random flight paths coalesce to form a seemingly haphazard swarm.

But the swarm you barged into is far more organized than you might expect.

In a study published June 25 in the journal Physical Review Letters, scientists show that swarming midges fleetingly dance with each other as if attached by a spring, giving researchers better insights into the rules underlying insect swarms and other animal collectives.

Nature is chock-full of wholes that are greater than the sums of their parts. Starlings self-assemble into teeming flocks; fish come together in schools that cut through the oceans; and trillions of individual cells interact to make up your body. These collective behaviors confound and inspire scientists and engineers aiming to better understand how medications traverse the body -- or how "swarms" of drones ought to fly best.

Studying these behaviors "is a truly of matter of aesthetic beauty," said Francesco Ginelli, a physicist at the University of Aberdeen in Scotland who was not involved in the research, but studies the mathematical patterns of animal groups. "If you have an eye for physics, you realize that [the groups] are not so different from other kinds of matter."

Studying collective groups of critters is tricky, according to Nicholas Ouellette, the senior author on the study and a physicist at Stanford University, in Palo Alto, California. "If you really want to understand this … you need to be able to measure what each animal is doing," he said. Historically, the field has depended on observations of animal behavior made with the naked eye, but recently, new technologies have allowed scientists to track hundreds of individual animals at a time as they form groups, finally giving researchers cold, hard data.

Previously, Ouellette had studied liquid turbulence with motion-tracking setups involving multiple cameras. Taking advantage of that experience, he and his colleagues began studying Chironomus riparius, a non-biting midge known for forming swarms of up to 100 individuals. After wrangling groups of midges into boxes in his laboratory, Ouellette used camera arrays to track individual insects, giving the researchers a chance to deduce the mathematical patterns behind the midges' motion.

Ouellette was curious to see if midges exhibited behavioral "effective forces," a powerful, simplified theory of swarm movement first proposed in the 1990s. According to the model, swarms stick together because individuals and their nearest neighbors get close, but not too close, coming together when far apart but avoiding collisions when near one another. This scheme works great in simulations: If you assign these attractive and repulsive "forces" to the members of a group -- and throw in some random motion for spice -- you get a jittery collective that convincingly looks like a swarm of midges.

But when Ouellette and his collaborators actually tracked the midges closely, they couldn't find much evidence that these effective forces existed, a result they published in the journal Scientific Reports last year. So what was really keeping the midge swarm together?

Ouellette and his colleagues realized they hadn't examined subtler, moment-to-moment changes, which would have required a different statistical approach. So they analyzed some  new midge swarms -- this time feeding the data through different statistical filters -- and found that the midges had two distinct modes of movement. One was exploratory, an essentially random flight pattern.

The other flight pattern was a little wackier. For only a few seconds at a time, individual pairs of midges would zigzag toward and away from each other, as if they were on opposite ends of a stretching and compressing spring between 25 to 75 millimeters (1-3 inches) long. Perhaps most surprisingly, the oscillating pairs of midges weren't always their closest neighbors. When locked in their springy flight pattern, other midges would often get closer, but the oscillating pair would ignore the would-be interlopers.

"Think about a crowded cocktail party: you're trying to have a conversation with somebody, and other people are moving through the room," said Ouellette. "They may come closer to you than the person you're talking to ... But you don't necessarily care about them as much," he said.

Midge swarms are made almost exclusively of males gathering to advertise themselves to females. Usually, female midges sit out on the sideline. Only when they want to mate do they fly into the maelstrom, to select a partner.

"It's kind of like a big singles bar," said Ouellette.

Ouellette and his colleagues suspect that the midges' short-term, springy interactions help midges identify their neighbor's sex -- a useful skill for a male trying to find a female if and when she arrives to the party.

Scientists know that female and male midges beat their wings at different rates. They think midges can use that difference to determine a neighbor's sex, by listening in on the wingbeats during the zigzagging flybys. There's some evidence to suggest the midges use this technique. When Ouellette played recordings of female midges' buzzing to groups of males, they wiggled in mid-air when flying by the speaker, as if they were trying to oscillate near -- and identify -- a female midge.

However, the researchers don't yet know if female midges also initiate the springy flight pattern, or exactly what happens when a pair stops zigzagging -- and either hooks up or breaks apart. The cameras the researchers used aren't high-resolution enough to thoroughly examine the subtler aeronautics of midge flirting.

Insect behavior experts were impressed with the study's use of unconventional tools. "I liked how [the study] started more at the physics end and ended up incorporating the biology," said Marla Sokolowski, a biologist at the University of Toronto who wasn't involved with the study. "I thought it was really, really interesting," she said.

The results also hit home that the effective force model is often too simplistic to accurately describe natural groups, which "is not totally unexpected," said Ginelli, given the complexity inherent in animal behaviors -- even those that we tend to think of as annoying.

Michael Greshko is a science writer based in Washington, D.C., who has written for NOVA Next, the National Academies, and NYTimes.com, among other outlets. He tweets at @michaelgreshko.