Watch: How Does a Dead Fish Swim Upstream?

Take a quick look at this trout swimming upstream. Notice anything unusual?

You’ve probably seen something similar countless times; the fish wriggles against the currents that push it backwards, slowly making headway until it turns and ducks out of the influence of the stream. Nothing special in that.

The only thing is, this particular fish is dead.

Yes, you read that right. No matter how lifelike it looks as it undulates across the tank, that same trout would just go belly-up if the current were switched off. So how can it possibly swim upstream?

A team of researchers from MIT and Harvard were equally surprised when they happened upon this phenomenon by accident. They’d been studying the way live trout conserve energy by swimming behind obstacles that block the current*, and unintentionally placed a dead fish in the experimental setup. When they took a closer look, they were stunned.

“It was incredible, very counterintuitive,” MIT researcher Michael Triantafyllou says, describing the shock he felt upon seeing the fish swimming upstream. He explains that while he knew trout were good at conserving and even extracting energy, he had no idea that they’d be able to extract enough energy from the surrounding fluid to swim upstream without expending any of their own energy. Immediately, the team started to investigate this new, seemingly impossible phenomenon.

As it turns out, objects that block the natural flow of water, like a rock or a boat, create a series of complex vortices in the current as the water navigates the obstacle. As anyone who’s tried to grab a fish knows, fish are quite flexible all down their spines, which allows the head and the tail to move independently of one another. In certain situations, the array of vortices forming behind an obstacle cause the body and tail to flap in resonance. This tilts the body in such a way that the vortices, which cause a pressure drop, apply a suction force that propels the fish forward.

As Triantafyllou explains, “You have a flow behind the obstacle, which creates a continuous stream of eddies. Each eddy contains energy and also causes the pressure in the fluid to drop… the eddy causes the body to flap back and forth, and the fish manages to extract energy.” Since all of the energy is supplied by the vortices, it doesn’t matter at all whether the fish is alive or dead, if the timing happens to be right.

In a world where we are always trying to improve efficiency, this discovery has obvious implications for aquatic vehicles. In fact, Triantafyllou says that a big limiting factor in our exploration of the oceans is the fact that the robotic devices that are often employed tend to have a very short battery life, usually just 8 hours. By creating new devices modeled after the trout, we could all but eliminate this constraint.

There is, of course, a catch. This group of researchers were quite lucky to happen upon the phenomenon, because in order for the fish to extract enough energy from the current to overcome its own drag, it must be very carefully positioned: too far away, and the vortices aren’t powerful enough; too close, and the suction from the obstacle pulls it in. When they set up the experiment with a dead trout, they accidentally hit the sweet spot, but, as Triantafyllou comments, “If you want to do it in real life, you cannot rely on accidents.” Instead, live fish have extraordinary sensors that allow them to create a map of the water flow so that they can adjust their positions to be as energetically favorable as possible.

Clearly, before the next generation of robots can begin their explorations, we’ll need to equip them with similar sensors. While the dead fish project was completed over a decade ago, Triantafyllou is currently working with partners Gabriel Weymouth (of the University of Southampton) and Jianmin Miao (of the Nanyang Technological University) to develop this technology—once more taking inspiration from nature.

One of their projects examines fish’s “lateral line”, an organ evidenced by a row of dotted scales along the fish’s sides that can detect the velocity and pressure of the water around it. Although the physical structure of the sensory units is reasonably well understood—simply put, they are bundles of hair cells encapsulated in a gelatinous material—it is extremely difficult to perform the calculations necessary to reconstruct a spatial map from each sensor’s input.

Another approach that shows promise comes instead from the harbor seal. These predators have extremely sensitive whiskers that can detect the disturbances left by potential prey up to 30 seconds after it passes by. It’s also able to distinguish the basic shape (square, triangle) of an object moving through the water. Triantafyllou’s group managed to 3D print “whiskers” modeled after the real ones, which have a “wavy” quality to them as they vary in diameter.

This work, which incorporates elements of biology, fluid mechanics, and engineering, is just one example of biomimetics, a field that has contributed innovations including Velcro and bird-safe glass. As Triantafyllou describes it, biomimetics is “learning from natural organisms for free to come up with some really ingenious solutions.” The idea is that since evolution is essentially a giant system for optimization—selecting the most efficient body design for swimming upstream, for example—we can take advantage of the engineering solutions that have already been developed through this process.  Clearly, this treasure trove of organic ingenuity is a tremendous natural resource that we’ve only begun to tap.

—Eleanor Hook

*Incidentally, it’s hypothesized that this is the reason fish swim in schools; those swimming in the wake of their companions expend much less energy than they would otherwise. For more on this, check out our 2015 podcast and post: “Flocks and Fluids“