Dr. Michael Ishida built a fish-like robot at the University of Cambridge that moves across a table with an awkward, undulating shuffle—and in doing so, solved one of evolution's oldest mysteries. The robot doesn't just mimic a single species; it reveals how unrelated fish scattered across the evolutionary tree have independently stumbled upon the same basic walking blueprint, one that may illuminate how vertebrates first conquered land hundreds of millions of years ago.
The breakthrough comes from a deceptively simple observation: several species of living fish—bichirs, lungfish, catfish, sculpin, and snakeheads—are capable of walking on land despite being far more efficient in water. For these species, the ability to escape onto shore when a predator approaches, or to shuffle between shallow-water environments like tide pools, is an evolutionary advantage worth keeping. But exactly how that walking ability emerges remained unclear, since individual walking fish species have been studied in isolation for years.
Ishida and his team, working across engineering, biology, and paleontology disciplines at Cambridge, changed that approach by looking for patterns that unite them. They began with a computer model based on observations of Polypterus senegalus, a gray bichir native to Africa, along with several other walking fish. What emerged surprised them: a recurring walking motion that appeared across multiple unrelated species. "A number of different fish, spread out across the evolutionary tree, and not closely related to each other, all do it," Ishida explained. "It's such a simple movement and can recur from a very basic starting point."
The researchers named this shared pattern the "undulating tripod gait." It looks inelegant—like a swimming fish that has been flopped onto land and is making do—but it is mechanically brilliant. The fish anchors its body using its front fins or head as a pivot point, then uses its tail to push the body forward around that anchor, essentially translating its swimming motion into a landward shuffle. What appears clumsy is actually one of life's most ancient solutions to a survival problem: escape predators or find new habitat without specialized limbs.
To test whether this pattern was truly optimal, Ishida and his colleagues built a physical robot fish and tried every variation they could imagine. Every alteration made it slower. "Any time we changed how the body bended, or what sequence it was bended in, it was worse," Ishida said. "It was surprising that the optimal walking pattern in the simulation and robot matched what the real fish actually do." The robot's most efficient movement closely matched both the computer model's predictions and the bichir's actual walking style.
This is the first time that unifying locomotive principles have been identified across multiple walking fish species—a striking example of convergent evolution, where different organisms independently evolve similar solutions to the same problem. The findings, published in Nature Communications, do more than explain modern fish. They offer researchers a framework for understanding one of the most significant transitions in the history of life: how vertebrates first moved from water onto land. Future work could be applied to fossil species like Tiktaalik, an important evolutionary bridge in that monumental shift. By understanding the mechanical simplicity of how modern fish walk, scientists are closer to understanding how their ancestors did it, too.