Bob Dylan asked: “Are birds free from the chains of the skyway?” Sure, this is a metaphor (in Dylan’s case, for a lost love) but it works because the complexities of avian flight—from migration and navigation to group dynamics—have long been a mystery, one with a preponderance of ideas, but few firm answers. Of course, insects and mammals also developed flight, in one of the great examples of convergent evolution, but bird flight in particular has long fascinated and mystified.
Homer and Aristotle both recorded details of bird migration; but human interest in bird flight probably goes back much further in time, if you go by indigenous people’s myths and legends. As long as there have been people watching birds, there have been theories as to how and why they do what they do. In the modern era, theories about why birds flock and why they migrate in v-formations have abounded, yet answers have been few. But new research using creative technology on both starling murmurations and bald ibis’ migration reveals that complex flight dynamics and rapid-fire adjustments based on sensory feedback previously believed impossible for birds are indeed occurring.
Built-in GPS? Not so fast
Bald ibises made the cover of the January 16, 2014 issue of Nature when researchers answered one of the biggest questions about how and why the birds migrate in a v-formation.
“For 50 or 60 years, there have been plenty of entirely theoretical papers which predict where birds put themselves in the v-formation,” says Dr. Steven Portugal, a postdoc in avian flight at University of London’s Royal Veterinary College (RVC). But nitty-gritty details were difficult to capture, and the data was incomplete: “Previously, people had used photos and videos, which doesn’t give the right accuracy. Those can distort height, if they were flying in the same plane, and it misses information. It only gives you snapshot of a flight as opposed to a dynamic image.”
Portugal’s team worked with the Structure and Motion Lab at the RVC on a built-from-scratch flight logger, which synced a 300-hertz accelerometer with a 5-hertz GPS and was light enough to attach to a bald ibis (it had to be less than 5 percent of the bird’s weight to ensure it wouldn’t unduly impact behavior). The combination of these two systems enabled the researchers to see exactly where the birds were and what their wings were doing by taking measurements at high frequency.
The team attached the loggers to 14 young ibises that were being reintroduced by a conservation group, the Waldrappteam, to their former range in Austria and Germany. The birds had to be taught their migration route and flew at intervals behind an ultralight plane. While most songbirds instinctively know migration routes, larger birds like pelicans, cranes, geese and ibises need to be taught where to go by their parents—or in this case, the conservationists. It was a unique opportunity to be able to track how the birds flew (and logistically far simpler than catching wild bald ibises).
The results from the loggers, taken over a 45-minute flight, revealed something previously believed, but never definitively proven: The birds timed their wing flaps and positioned themselves relative to other birds to maximize efficiency. “The main finding was not only that they position themselves in the best possible position for catching updraft, but that it’s an active process. The wingtip of a bird following another takes the same path as the one in front, so they adapt when and how they flap to capture as much upwash as they can,” says Portugal.
Engineering for the birds
Dr. George Young is a mechanical engineer who has closely examined starling murmurations (those giant flocks of coruscating birds that viral videos are made of) to understand optimal group behaviors—and not just to answer questions about how and why the small birds flock as they do (interesting enough), but because that information could be useful to his work in designing non-biological intelligence. “We’re looking at how we can design groups of sensors or robots to allow them to do something sophisticated and intelligent for cheap,” says Young.
Starlings could provide the answers Young needs, since they have solved the problem of communicating over large groups with plenty of (information) noise. Young calls this kind of work, “bio-inspired engineering” – using nature’s solutions to solve ongoing quandaries.
While it was known that in starling flocks each bird pays attention to its seven closest neighbors, what wasn’t understood was why. In a January 2013 research paper in PLOS Computational Biology Young and Naomi Leonard, his PhD advisor at Princeton (where he was pursing his doctorate), as well as colleagues from Sapienza University in Rome, determined seven is the number that “optimizes the balance between group cohesiveness and individual effort,” according to the paper.
“Essentially, if the bird is paying attention to too few neighbors you can’t pass information through the whole flock. If a bird is watching too many, they’re not getting more information, they’re just paying the cost. Six or seven is the minimum number of neighbors you need in a group to keep the flock connected,” says Young.
To figure out the magic number, Young used video data that was analyzed frame-by-frame. The position and velocities for each bird in the group were tracked and plotted. “We took that position data and used that to reconstruct hypothetical interaction networks,” says Young. And then some creative use of technology came into play.
MatLab, a program widely used in engineering for stress analysis and fluid dynamics, was familiar to Young as an engineer, but in his starling research he used it to simulate biological systems instead. “We ended up working on a very large matrix, with a row and column for each bird, but MapLab has its own programming language that allows you to quickly code up large computations” like those coming from hundreds of starlings in a murmuration, says Young.
By using an engineering program for biological analysis, the magic number was found, and Young hopes to apply this knowledge to other systems that are dealing with signals and noise, as starlings do so effectively. “The fact that the same number of neighbors is optimal over a range of flock sizes and densities—as well as, to a certain extent, typical flock thicknesses—suggests that the number of neighbors that a bird interacts with could be an evolved trait,” says Young.
Speaking of evolution, what could be a more perfect example of where we are now than an engineer using software to understand a natural solution to a problem that will eventually make smarter robots?