The Physics of Animal Motion

By Ernest Davis

How to Walk on Water and Climb up Walls: Animal Movements and the Robots of the Future. By David Hu. Courtesy of Princeton University Press.

Animals walk, run, swim, fly, glide, hover, slither, burrow, and swarm. They also regularly move material in and out of their bodies. How to Walk on Water and Climb up Walls, a new book by biophysicist David Hu of the Georgia Institute of Technology (Georgia Tech), describes the fascinating science behind the physical and biological principles of animal motion. Hu then links these actions to the engineering of robots that move in similar ways.

The book opens with Jerry, Hu’s girlfriend’s toy poodle, shaking himself dry after a bath. Jerry shakes about seven times per second and his shaking generates forces up to 12 times gravity; in a fraction of a second, he can eliminate 70 percent of the water in his fur. Hu—with help from his student, Andrew Dickerson—built a wet dog simulator that spun a clipping of Jerry’s hair. At both Georgia Tech labs and Zoo Atlanta, the pair took high-speed films of creatures ranging from bears (which shake four times per second) to mice (which shake 29 times per second).

Evolution has managed to design animals that exploit the subtlest properties of the media through which they move. Water striders, as their name suggests, are small insects that walk on water. They are light enough—and their feet are long enough—for the surface tension to support them. If you blow on them gently, they simply glide along the water’s surface. Understanding how they move and from where the traction comes was an enigma. Stanford University biophysicist Mark Denny discovered that adult water striders create tiny waves in the water that provide the necessary propulsion. But the legs of infant insects are too small to generate the necessary waves. The mystery of how they move is thus known among water-strider aficionados as “Denny’s Paradox.” As a project for a fluid mechanics course, Hu determined that another mechanism is involved; the infant water striders’ legs produce a tiny vortex in the water, the momentum of which is enough to push the insects forward. Armed with this knowledge, mechanical engineer Brian Chan built a water-strider robot¹ from the aluminum in a soda can. He powered his robot with a thread from an athletic sock, which is also apparently a marvel of modern material design.

Hu proceeds to describe his work—and that of his students, teachers, and collaborators around the world—with creatures of all kinds. He explains how snakes slither across a flat surface, how worms tunnel through mud, and how common sandfish (a type of lizard) burrow through sand. He details the flight mechanism in flying snakes and the altogether different flight apparatus in bumblebees. Hu explores how armies of ants utilize their own bodies to build bridges over gaps, and informs readers of the simple distributed feedback scheme that ants use to decide on bridge placement. He even conducts comparative studies of urination times in different creatures and finds that they are remarkably constant — 10 to 30 seconds in a variety of animals, ranging from his infant son to elephants. Hu and his colleagues investigate the impact of eyelash length in clearing water from the eye (short lashes are much more effective than long lashes), and the role of shark scales in dictating swimming patterns. They analyze the creation of different locomotion mechanisms that require the least energy, and research how—in the right kind of current—a dead fish can “swim” using no energy at all.

These experiments involve both close interaction with all kinds of animals and a wide range of ingenious experimental devices and analytical techniques. Exploring worm motion requires a type of transparent mud, and studying swimming creatures calls for dyed water. Hu and his colleagues utilize high-speed cameras and delicate motion detectors. They employ state-of-the-art computer simulation techniques to calculate the theoretical predictions of fluid and material mechanics, and compare these results to experimental findings. If a scientist can then build a physical robot using the same principles, they would have tangible confirmation of the theory’s accuracy (plus a very cool robot). Hu’s writing is chatty, entertaining, gracious, and very clear. The book itself is amply illustrated with line drawings and a dozen color plates.

I find it hard to imagine a direction of scientific inquiry with more obvious natural appeal to the general public; Hu’s work perfectly fits the “Science is interesting and fun!” campaign that scientists eternally wage. I was therefore shocked—though perhaps in hindsight I should not have been—to learn that Hu’s research has been singled out for mockery by the professional obscurantists who infest broadcast news and politics. Why, they cry, should public funds be spent calculating how long it takes elephants to pee? In 2016, Fox and Friends featured three of Hu’s projects on their “Wheel of Waste.” 

The very fact that this kind of work is so easily understood can become a club to beat it with (most scientific research is protected from this type of attack by the sheer difficulty of explaining what it is about). Even more depressing is the fact that—as Hu discusses towards the end of How to Walk on Water and Climb up Walls—scientists working in these areas often struggle to get institutional support at universities or research labs for this type of interdisciplinary work. Perhaps for this reason, Hu’s last chapter is somewhat defensive in tone, laying out the potential impact of these studies for both medicine and robotics of all kinds. On the one hand, an enhanced understanding of human motion can lead to better diagnostics (for example, physicians can diagnose Parkinson’s disease and similar ailments in part by measuring gait) and improved assistive devices, exercise apparatuses, and prosthetics. On the other hand, the best robot for many tasks is not one that walks like a person, but rather moves in one of the many ways that animals do.

¹ In this book, the word “robot” simply means a mechanical toy; there is no requirement that it must be controlled by a computer.

Ernest Davis is a professor of computer science at New York University’s Courant Institute of Mathematical Sciences. His book with Gary Marcus, Rebooting AI: Building Artificial Intelligence We Can Trust, was published in September.