Squirrels show off killer parkour moves as they leap from branch to branch

To jump or not to jump: that is the question —

Pushing off vertical surface helped squirrels adjust their speed for a better landing.


High-speed video of squirrels executing amazing leaps in pursuit of peanuts captures the animals’ innovative solutions to reaching their goal. Video courtesy of Roxanne Makasdjian and Jeremy Snowden, University of California, Berkeley.

The campus of the University of California, Berkeley, boasts three known wild populations of fox squirrels, who spend their days cavorting among the tree branches and foraging for sustenance. They’ve also now made a contribution to science, thanks to a series of experiments by UC-Berkeley researchers aimed at assessing how squirrels figure out whether or not to leap from one given tree branch to another.

In the process, the team caught several squirrels resorting to innovative moves, reminiscent of parkour, to execute especially tricky maneuvers, re-orienting their bodies to push off a vertical surface to ensure a smooth, safe landing. The team described its findings in a new paper published in the journal Science.

Squirrels are masters at navigating through the treetops, jumping from branch to branch without falling. “As a model organism to understand the biological limits of balance and agility, I would argue that squirrels are second to none,” said co-author Nathaniel Hunt, a former UC-Berkeley graduate student who is now researching biomechanics at the University of Nebraska, Omaha. “If we try to understand how squirrels do this, then we may discover general principles of high performance locomotion in the canopy and other complex terrains that apply to the movements of other animals and robots.”

The UC Berkeley team thought that the key might be trial-and-error learning and constructed an outdoor obstacle course in a small forest area located on the west side of campus (the Eucalyptus Grove/Grinnell Natural Area). “Squirrels from this population were reliably found in the study area daily,” the authors wrote in their supplementary material. They used a syringe to mark individual squirrels with black fur dye in different patterns to keep track of them.

The experimental apparatus consisted of a ramp leading to launching platforms of varying flexibility and rigidity—basically rods that simulated branches, magnetically attached to a vertical steel wall—and a landing perch opposite (a dowel wrapped in athletic tape), with varying distances between them. There was also a cup on the landing perch holding a peanut to motivate the squirrels to make the jumps between the launching platform and the landing perch.

The experiments were preceded by a training period in which small pieces of peanut were placed on the ramp leading to the jumping platform to lure hungry squirrels, with a half-peanut in a cup at the end of the beam as a reward. Once the squirrels were comfortable getting the peanut from the cup, the team started gradually increasing the gap between the end of the beam and the landing platform. Most of the squirrels mastered this task within 30 minutes.

  • Fox squirrel will jump for peanuts. Several populations of fox squirrels call the UC Berkeley campus home.


    Lawrence Wang, UC Berkeley

  • Fox squirrel, apparatus, and experimenters on UC Berkeley campus. Left to right, Nate Hunt, Judy Jinn, Lucia Jacobs, and Aaron Teixeira.


    Gregory Cowley

  • Aerial and landing postures as squirrels learned to leap from an unfamiliar flexible beam.


    N.H. Hunt et al., Science 2021

  • Four landing maneuvers observed in the first two experiments.


    N.H. Hunt et al., Science 2021

  • A squirrel demonstrates innovative parkour-like skills as it leaps across the gap to get that tasty peanut reward.


    N.H. Hunt et al., Science 2021

  • Sequence of cyan dots represents center of mass at each video frame during the aerial phase of the leap.


    N.H. Hunt et al., Science 2021

  • (C) Still frame of a squirrel landing with a small landing error. (D) Still frame of a squirrel landing with a large landing error. No squirrels in the trials actually ever fell.


    N.H. Hunt et al., Science 2021

The first experiment focused on decision-making. The team used three “branches” of varying flexibility attached to the jumping platform: a birch rod, a plastic tube, and a plastic tube covering a brass rod, all wrapped in athletic tape so that the squirrels had consistent traction. Varying the flexibility of the perches was crucial to the experiment, since it forced the participating squirrels to figure out a workable tradeoff between stability and leaping distance.

The UC Berkeley researchers made things a little bit harder for the squirrels in the second experiment, increasing both the range of flexibility or rigidity of the branches and the minimum gap distance. They used a rigid beam as a control.

The squirrels figured out how to modify the biomechanics of their jumps quite quickly. They were a bit more cautious when it came to jumping from more flexible fake branches, and it took them a couple of tries to master it. “This behavioral flexibility that adapts to the mechanics and geometry of leaping and landing structures is important to accurately leaping across a gap to land on a small target,” said Hunt.

Hunt et al. observed several distinct landing maneuvers that the squirrels used to compensate for leaps that were too fast or too slow. For example, they would roll forward around the branch if they overshot the landing platform when they jumped and would grab the branch with their front legs and swing under it if their jump fell a bit short, pulling themselves up to get the peanut. And sometimes the squirrels executed the leaps just right to achieve a direct landing on the perch.

Ultimately, the team found that the flexibility or rigidity of the landing platform was the most critical factor in whether the squirrels decided to jump, much more so than the distance of the gap. And none of the squirrels ever fell, perhaps because they have such sharp claws.

“They’re not always going to have their best performance—they just have to be good enough,” said Hunt. “They have redundancy. So, if they miss, they don’t hit their center of mass right on the landing perch, they’re amazing at being able to grab onto it. They’ll swing underneath, they’ll swing over the top. They just don’t fall. This combination of adaptive planning behaviors, learning control, and reactive stabilizing maneuvers helps them move quickly through the branches without falling.”

The researchers were surprised to note that squirrels sometimes resorted to innovative, parkour-like moves to close the gap and get that tasty peanut, reorienting their bodies mid-leap so that they could push off the vertical wall, thereby adjusting their speed for a smoother landing. To test how often the squirrels resorted to this strategy, the team switched up both the gap distance and landing height. All told, ten squirrels completed a total of 324 parkour leaping trials. The team found that the squirrels used this maneuver pretty consistently for medium and long jumps, but never for the shorter jumps. The height of the landing platform wasn’t a relevant factor in those decisions.

“Just as movement in the real world requires flexibility and creativity, researchers studying natural locomotion must be as ingenious as their animal subjects,” Karen Adolph (New York University) and Jesse Young (Northeast Ohio Medical University) wrote in an accompanying perspective. “The trick is to capture movement in all its complexity while retaining sufficient experimental control and measurement fidelity. The study of Hunt et al. is a beautiful example. Their unexpected results elucidate what every homeowner knows: squirrels are clever acrobats when navigating complex environments.”

The UC-Berkeley team will continue to study the intricacies of squirrel biomechanics and how it links to cognition in hopes of one day building a robot with similar leaping capabilities. “I see this as the next frontier: How are the decisions of movement shaped by our body?” said co-author Robert Full. “This is made far more challenging, because you also must assess your environment. That’s an important fundamental biology question. Fortunately, now we can understand how to embody control and explain innovation by creating physical models, like the most agile smart robots ever built.”

Main paper: DOI: Science, 2021. 10.1126/science.abe5753  (About DOIs).

Related Perspective: DOI: Science, 2021. 10.1126/science.abe6733  (About DOIs).

Listing image by Judy Jinn, UC Berkeley

Read More

Jennifer Ouellette