When animals swarm they exhibit a complex collective intelligence that could help us build robots, heal wounds and understand the brain
IAIN COUZIN does not have fond memories of field research. Early in his career, he travelled to Mauritania in north-west Africa to follow a swarm of locusts. Devastation caused by the insects meant no one was selling food and the team was forced to live off dried camel entrails. Couzin, a vegetarian at the time, was violently ill. "I was hallucinating – I thought I was going to die." By the time he recovered, a huge sand storm had blown in. The researchers were trapped in their tents for several days and when, eventually, they emerged, the locusts had gone, blown away by the storm. "I was out there for two months and I got absolutely no usable data," he says. "It was the worst experience of my life."
Fieldwork can be difficult at the best of times, and it would appear that Couzin, who is at Princeton University, is not the only swarm scientist averse to it. One of the tricky things is how to study the interactions between animals when their numbers are so huge. So researchers have generally stayed indoors with their computer models. However, these are only as good as the information you put into them, and often they have not proved terribly enlightening. You can recreate swarm-like behaviour without really understanding why it exists. Now, though, researchers are starting to see swarms as living entities with senses, motivations and evolved behaviour. From this new view is coming a much better understanding of how animals act collectively.
This does not simply tell us about flocking birds, shoaling fish, swarming locusts, and the like. It has implications for how we understand all sorts of collective action. There is a limit to what a single organism can compute, but the combined information-processing power of a swarm is more than the sum of its parts. Applying this concept to other complex systems provides insights in all sorts of areas, from fighting disease to building robot swarms. It might even provide a way of thinking about the human brain.
For a long time, the standard approach to studying synchronised movement was to model the animals concerned as "self-propelled particles" following a few simple rules, such as "keep a body length away from your nearest neighbours" and "match the speed and orientation of the organism in front". This physics-led approach, which treats animals as mindless objects, is almost certainly too simplistic – a point that was brought home to Couzin a few years ago.
In an attempt to understand how locust swarms march together across an area of land, he and his colleagues had built a model which represented the insects as a collection of particles, rather like the atoms in a gas. To coordinate movement and prevent collisions, each "particle" simply had to adjust its speed and direction in response to the speed, proximity and direction of its neighbours. The team's findings were published in Science in 2006. Only later did they discover the flaw in their model. Watching real locusts in the lab, they were surprised to find fewer at the end of their experiments than at the start. Far from avoiding collision, they were exterminating one another as they marched. "We discovered by chance that the swarm is driven by cannibalism. Everyone is trying to eat everyone else while avoiding being eaten," says Couzin. "That was a real wake-up call."
Since then, Couzin and his collaborators have seen swarming in a different light. "This isn't just about physics," he says. "These are biological organisms: they're responding to sensory information." Understanding this makes studying swarms more challenging because you need to consider the capabilities and motivations of their members. But with the help of new technology, this is exactly what Couzin and others are doing and, in the process, overturning some preconceived ideas about swarms.
Info in flow
Take shoaling fish. Olav Handegard, who works in Couzin's lab and also at the Institute of Marine Research in Bergen, Norway, is using sonar imaging to reveal what is going on in the murky waters of Louisiana's estuaries when shoals of Gulf menhaden come under attack from spotted sea trout. Like many schooling fish, they split up into smaller pods, which according to received wisdom is a way of evading predators. Not so. Handegard has found that this is what the trout are aiming for: they do their best to break up the menhaden shoal because it is easier to take a fish from a smaller group. For the menhaden, the intact shoal is the best place to be because news of a predator's presence reaches them more rapidly in a large shoal. Each fish reacts to the movements of its nearest neighbours to create a "wave of turning" that propagates 15 times as fast as a fish can swim, and faster than the predator too. The more eyes there are to spot danger and the more neighbours' movements there are to follow, the better the information flow.
- New Scientist
- Not just a website!
- Subscribe to New Scientist and get:
- New Scientist magazine delivered every week
- Unlimited online access to articles from over 500 back issues
- Subscribe Now and Save
If you would like to reuse any content from New Scientist, either in print or online, please contact the syndication department first for permission. New Scientist does not own rights to photos, but there are a variety of licensing options available for use of articles and graphics we own the copyright to.
