How three-legged dogs actually prefer to move

Doges Editorial · 2026-06-01 · 5 min read

A Royal Veterinary College study placed 12 canine amputees in a motion-capture lab and found something counterintuitive: slow walking is actually harder for three-legged dogs than a canter. The findings carry immediate implications for post-surgery recovery.

Dr Zoe Davies was sharing a postdoctoral office at the Royal Veterinary College when she started watching her officemate's dog more carefully than she watched most dogs. The dog had three legs. Davies had spent her career studying how birds move, how fish move, the elegant mathematics of locomotion in creatures built to specification. She had never worked with dogs. She started thinking about it.

Inside the Structure and Motion Laboratory

The resulting study, published in the Proceedings of the Royal Society and covered by Vet Times on 29 May 2026, is the most detailed analysis of three-legged dog movement to date. Davies and her colleagues — Professor Jim Usherwood and RVC undergraduate Aimee Savage — recruited 12 canine amputees from across the UK via RVC's social media channels: six dogs missing a forelimb, six missing a hindlimb.

Each dog walked, trotted, and ran through RVC's Structure and Motion Laboratory along a 10-metre track. A 3D motion capture system tracked precise body movements in real time. Force plates embedded in the floor measured how much weight each remaining limb absorbed on each contact. By the end of data collection, the researchers had analyzed 825 individual limb-stance events — enough to draw conclusions with confidence.

What happens at speed

The first finding was almost reassuring. At higher speeds, three-legged dogs adopt a gait that closely resembles the gallop of a four-legged dog. The mechanics are different — there's a limb missing — but the pattern is recognizable, the rhythm consistent. As the paper puts it: "none of the gaits needs to be considered random." These dogs are not improvising chaotically. They have found a solution, and they run it reliably.

For the six dogs missing a forelimb, that single remaining front leg absorbs approximately 50 percent of the dog's total bodyweight on each landing. For hindlimb amputees, the load is distributed more evenly across the three remaining legs. Understanding where the stress falls matters for thinking about long-term joint health and which limbs are candidates for early wear.

The surprise about slow walking

The more unexpected finding emerged at lower speeds. When dogs slowed down, the data revealed two strategies. Some used a slowed-down version of the gallop. Others adopted what the researchers called "uncoupled walking" — a gait in which the two-limb side moved in a walking rhythm while the single remaining limb contacted the ground more than once per stride. Both strategies, the data suggested, were less comfortable than moving faster.

It was apparent that this was not a preferred speed for any of the dogs and appeared laborious. Given that there were distinct alternative gait strategies at slow speeds, this could indicate that such speeds were uncomfortable, similar to those around the human walk–run transition speed. It is worth considering this during post-operative rehabilitation, and potentially not walking dogs too slowly, but aiming for a more comfortable middle speed.

— Davies, Savage, Usherwood — Proceedings of the Royal Society B, April 2026

The human analogy

The comparison to human gait is precise. Humans have a threshold speed just below jogging where walking becomes mechanically inefficient — the body wants to run, and being asked to walk at that speed feels wrong in a way that's hard to explain. The three-legged dogs in the study showed a structurally equivalent phenomenon: a speed range where the available gaits are laborious, and going a little faster would be easier.

The clinical implication is immediate. After a dog loses a limb — often following osteosarcoma, or severe traumatic injury — the recovery protocol typically involves very short, very slow walks to protect the surgical site. That remains appropriate in the early days. But as rehabilitation progresses, the data now supports what some owners and physiotherapists had noticed intuitively: the dog is more comfortable at a trot than at a shuffle.

How the study started

The focus of my postdoctoral research was in two-legged and four-legged locomotion, but I was inspired to look at three-legged locomotion when I started sharing an office with a canine amputee. The great thing about research is that you can often end up pursuing avenues you didn't anticipate at the start.

— Dr Zoe Davies, lead author, now Lecturer in Veterinary Sciences at Harper & Keele Veterinary School

The Wellcome Trust funded the work. The 12 dogs were recruited via social media, traveled to the RVC's London campus, did their runs across the 10-metre force-plate track, and went home. They are presumably still moving at a comfortable middle speed somewhere in the UK, their owners perhaps noticing that the trot has always looked more natural than the slow walk.

Beyond veterinary medicine

There are an estimated 1.5 million dogs living with a limb amputation in the United States alone. For most of them, life continues at a quality their owners describe as unexpectedly high. Studies consistently show that owners report far greater satisfaction with their dog's post-amputation life than they predicted before the surgery. Dogs adapt faster than the humans caring for them expect. Research like this turns that adaptation from an impression into a measurable fact.

The authors also note that the findings could help inform adaptive legged robotics — machines designed to maintain function after a component fails. How a dog reorganizes its gait when a leg is gone turns out to be a useful model for thinking about locomotion under constraint. But that application is downstream.

The immediate application is the walk. If your dog has three legs, or if you have ever watched a tripawd and wondered whether you were slowing down out of care or out of something else, the motion-capture data has an answer: let them move at a pace that feels like theirs. At a trot, their gait is not a compromise. It is a solution they worked out on their own.