Automatic Transmission Gears Found in Human Legs

A new paper uses the word “gears”
to describe how legs mechanically
shift when speed increases

 

An astonishing case of “nature had it first” is found in a new paper in the Proceedings of the National Academy of Sciences. Want to learn about a continuously adjusting automatic transmission that shifts gears as speed increases? No need to take a course in physics or automotive technology. Look no further than your own legs and ankles.

Shifting gears to power human walking (Martin et al., PNAS, 24 Aug 2026, online 8 Sept 2026). Three mechanical engineers from the University of Wisconsin-Madison and a colleague from the University of Alberta describe the human ankle as the place where a wonder of biophysics takes place.

Human muscles outnumber the minimum needed to actuate our joints, yet the advantage of this redundancy remains unclear. We show that redundancy provides a system of biomechanical “gears” that allow one to effectively shift power production at the ankle between muscles during human walking. This finding offers a compelling perspective on muscle anatomy as a possible adaptation for versatile performance, with implications for understanding locomotor adaptability, improving rehabilitation, and advancing assistive and bioinspired technologies.

If human muscles outnumber what’s needed to activate our joints, then we’re looking at a case of “overdesign” in the human body. And since humans have centuries of experience with levers, gears, moment arms and the concept of mechanical advantage, we can look at our anatomy and physiology afresh to see how the body solves physics problems. The problem in this case involves maintaining mechanical advantage when speeding up from walking to jogging to sprinting.

Note: To make sure I didn’t misconstrue what these authors were saying about “gears” in the paper, I asked Gemini for help (see dialog at end of this article).*

Before enjoying this remarkable paper, let’s de-Darwinize it by getting rid of the useless Darwinese popping up here and there:

• “Human evolution has led to anatomical adaptations that enable efficient, robust, and adaptable movement.” Evolution is not an intelligent designer. How could the Stuff Happens Law ‘lead to’ efficient, robust, and adaptable engineering solutions?
• “One intriguing evolutionary feature is muscle redundancy, which is the presence of more muscles than degrees of freedom in the human body.” That’s overdesign, not evolution. Why call this an ‘evolutionary’ feature and not a design feature?
• “This gear-like behavior provides a potential explanation for the evolutionary persistence of redundant muscle systems and underscores the importance of having a range of moment arms in human motor control.” If it is only a ‘potential’ explanation it is not yet an explanation at all. Natural selection is not going to maintain redundancy without function.
• “Evolutionary changes in human plantarflexor moment-arm geometry may reflect a functional specialization for sustained power generation in bipedal walking compared to other primates.” Begging the question. Chimps are chimps, and humans are humans. Knuckle-walking and upright bipedal walking have different mechanical requirements. Why connect distant dots, and not instead conclude that both are intelligently designed for their needs? The only reason is a philosophical bias for materialism.
• “This behavior has analogies to gearing in mechanical systems, and likely reflects evolved anatomical features and behavior that enhance the energetics of long-duration locomotion.” Likely only to a DODO bigot. Engineers know design when they see it.
• “This general insight enhances our understanding of muscle redundancy as it relates to locomotion and evolution, while also having practical relevance for assessing the causes and implications of a variety of musculoskeletal gait disorders and injuries.” The understanding of locomotion belongs to physicists. The understanding of redundancy belongs to engineers. The assessing of causes of injury belongs to medicine. Evolution is a useless ball and chain in this sentence, serving no function but to dull the mind from the wonder of functional design in humans.

Now to the science. Here’s the upshot of what these mechanical engineers found in our legs:

In this study, we investigate whether muscle redundancy enhances locomotor efficiency by providing an array of moment arms that function like mechanical gears. Like an engine shifts gears for efficiency—using larger gears for acceleration and smaller gears for sustained higher speeds—we hypothesize that humans shift gears by phasing power generation across muscles with different moment arms during locomotion.

So that’s why we need redundant muscles for locomotion. Different speeds expand or reduce the moment arm (i.e., leverage distance; how far away from a pivot point a force is pulling or pushing), which changes the efficiency of the power invested for movement.

Think of the old 10-speed bicycles with five rear gears and two front gears. By operating a lever on the handlebar, the rider can make the chain jump from a large gear to a smaller one for efficiency when pedaling uphill, or to a larger one on flat ground for going fast. For the same radius of pedaling circle, the bicyclist changes the gear ratio, and thus the moment arm. She gets the most efficient locomotion depending on the need.

What’s neat about the body’s solution is that everything is automatic. Decide to slow down or speed up, and the body does the work for you. It figures out what muscles need to be activated to “switch gears” for the best output from power invested. You don’t even need to think about it. The authors referred to bicycling in this quote:

Gearing plays a role in many aspects of locomotion. At the whole body level, progression of the center-of-pressure under the foot scales the internal moment required of the ankle plantarflexors. At an individual muscle level, the moment arm of a muscle-tendon unit about a joint modulates both the moment-generating capacity and the speed of muscle shortening. This tradeoff between isometric strength and speed of contraction has been linked to performance in tasks such as pedaling and sprinting. Within the muscle itself, the angle between muscle fascicles and the line of action of the muscle-tendon unit determines the gear ratio between the shortening velocity of the muscle fibers and the muscle-tendon unit. Each of these manifestations of gearing have been shown to influence the mechanics and shortening velocities of individual muscles, in turn playing an important role in energetics and efficiency.

The ankle plays a role in lengthening or reducing the moment arm.

While the architecture and absolute moment arm lengths distinguish the TS [triceps surae] from the invertor and evertor muscles, their mechanical advantage is also influenced by how moment arms change with ankle joint posture. As the ankle plantarflexes, both muscle groups exhibit increases in moment arm length due to changes in muscle paths and shifts in rotation centers from joint geometry. However, the invertors and evertors show a greater relative increase in moment arms, on average 50% (25 to 80%, 6.5 mm) compared to 9% (5 to 15%, 4.6 mm) for the TS. The increase in moment arms of the OPF [other plantarflexors] with ankle plantarflexion is well aligned with increasing contributions to plantarflexion in late pushoff where ankle plantarflexion rapidly increases.

The “automatic transmission” between ankle and muscles involves tradeoffs that, in combination, contribute to smooth, continuous locomotion. Unlike a car with manual or automatic transmission where you can feel the gearshift, there is no jump between gears in the leg-gear system.

We argue in our analysis that smaller plantarflexion moment arms are beneficial at higher velocity, but this effect is bounded by requirements for moment generation, i.e., if moment arms become too small, moment generating capacity becomes a limiting factor for power production. This is in contrast with the muscle contraction velocity limitations associated with large moment arms. Thus, the important result of the observed moment arm changes across muscles with plantarflexion may actually be a reduction in the gap between gear ratios across muscles (SI Appendix, Fig. S1), which may be beneficial for smooth phasing of power from the TS muscles to the OPF. The effect of joint-angle-dependent moment arms on gear ratio spacing suggests that humans may have developed muscles and joint shapes to leverage power phasing between muscles.

I don’t know if you “developed” this highly efficient and functional system yourself. I sure didn’t. But we can all enjoy what our Creator gave us – a smooth, overdesigned system for walking and running that shifts gears smoothly between speeds, maximizing the power put into the muscles.

A surprising aspect of this study is the relatively large magnitude of work contributions attributable to the plantarflexor muscles with smaller moment arms. By combining tensiometry and motion analysis, this study provided estimates of mechanical energy contributions among muscle groups. Our experimental results suggest that the other plantarflexor muscles produce peak moments 21 to 31% as large as the peak net plantarflexion moment and contribute 40 to 45% of the net positive work during pushoff. These contributions may be larger than anticipated due to their previously overlooked capacity to sustain force at high plantarflexion velocities.

Who anticipated smaller contributions during pushoff? Who overlooked the capacity to sustain force? Was it creationists or evolutionists?

The authors remind us that walking involves discrete steps. Each step involves a landing, stabilization, and pushoff. But even within our step-delimited locomotion, a walk in the park can seem very smooth. Thank the contributions of multiple coordinating parts in the legs, ankles, feet, and toes for that pleasure.

The simulations suggest systematic phasing of muscle contributions to the net ankle power trajectory. At the beginning of pushoff, the TS are the primary contributors to net power, with the power arising from both Achilles tendon recoil and rapid shortening of the gastrocnemius [calf muscle] and soleus [in front of the calf] muscle fibers. In contrast, force levels of the OPF are sustained, delaying tendon recoil until later in pushoff. As a result, the power generation just prior to toe-off arises primarily from these muscles with smaller moment arms, notably the posterior tibialis, peroneals, and extrinsic toe flexors (FHL, FDL).

In the conclusion, the authors revisit the analogy of gear shifting, pointing out similarities and differences. Which one sounds better designed?

Gear shifting within the human musculoskeletal system has both parallels with and distinctions from traditional mechanical systems. Like most actuators, muscle-tendon units operate most effectively within specific velocity ranges. However, while mechanical systems typically shift gears to allow a single actuator to operate at an optimal speed, humans shift gears by phasing power generation between actuators with different moment arms.

The authors think that robot designers can learn from human anatomy. Humanoid robots are doing pretty well these days, and dog-like robots can help humans on the battlefield or in rescue operations. But we still must marvel at the real natural leg:

Interestingly, the importance of redundant actuators in human motion is in contrast with what is typically employed in robotics, where a one-to-one pairing of actuators to degrees of freedom is typical for simplicity in design and control systems. These systems can encounter difficulties in applications with both high torque and high velocity demands, so a more biomimetic approach may be warranted depending on the application.

Oh, but our Creator didn’t shortchange our design for simplicity. He gave us the deluxe model.

• See also Lieberman’s study on all the anatomical contributions for distance running here (27 June 2019).
• See this anatomist’s statement about the design of the human foot (27 June 2019).

This was a fascinating paper. Without dispute, evolutionists can get some really good science published. But the evolution part contributes nothing to the science. It’s like the flag on the handlebar, or the horn, that makes others look and hear the cyclist as he passes by. Darwinese doesn’t help him get up the hill faster. It’s just dead weight to satisfy the judges.

---

*Gemini Dialogue

DFC: Please evaluate the following paper in layman terms, especially clarifying what they mean by gears.

Gemini: The 2026 PNAS study by Jack A. Martin et al. demonstrates that the human ankle functions as a biological mechanical gearbox during walking by distributing work across muscles with varying moment arms. The authors define “gears” through the geometric layout of the muscle tendons: larger moment arms act as a “low gear” to generate massive torque during slow-rotation push-off phases, while smaller moment arms act as a “high gear” to maintain force and efficiency during high-speed joint movements.

DFC: Thank you. Then where is the “stickshift” to switch gears, and how does a runner operate it?

Gemini: This biological system lacks a physical stickshift, functioning instead as a passive automatic transmission where the runner does not need to consciously change gears. The transition occurs automatically through muscle-tendon dynamics and load redistribution as pace accelerates, allowing the runner to operate the system simply by changing their speed or pushing harder.

DFC: Got it. So we can say metaphorically that the human ankle has automatic transmission. Is there a set speed where the shift occurs, or is it continuous?

Gemini: The transition in the human ankle is a continuous, fluid process driven by the changing velocity of the joint during push-off, rather than a discrete shift at a specific speed. The ankle functions analogously to a continuously variable transmission, seamlessly blending muscle workloads across overlapping groups to sustain momentum as running or walking speed increases.

DFC: I’m reading the paper, but just want to accurately construe the metaphor of “gears” that the authors use. Obviously the human leg and ankle differ substantially from a mechanical gearbox or transmission. You mention overlapping groups. How many of those are there?

Gemini: In the study, the authors categorize the ankle’s “gears” by looking at two primary functional divisions of muscle groups acting across the joint:

• The “Low Gear” Group (The Triceps Surae): This is the large, primary muscle group at the back of your calf (the gastrocnemius and soleus), which connects to the prominent Achilles tendon. It has a large moment arm. In mechanical terms, it acts as a low gear because its attachment point sits further from the joint center, giving it massive mechanical leverage to generate high force and torque during the initial, slower phase of your foot’s push-off.
• The “High Gear” Group (Deep/Smaller Flexors): This comprises a collection of smaller, deeper muscles including the posterior tibialis, the peroneals, and the extrinsic toe flexors. These muscles have small moment arms because they wrap tighter and closer to the center of the ankle joint. Like a high gear on a bicycle, they trade away brute torque leverage in exchange for velocity. As the ankle begins spinning faster and faster near the end of a stride, these smaller muscles are structurally primed to maintain high-speed power production when the massive calf muscles begin to lose their mechanical advantage.

Rather than having dozens of individual gears, the ankle relies on this elegant two-tiered layout. The transition between them is seamless: as joint speed picks up during push-off, the workload naturally cascades and distributes from the heavy low-gear calf muscles to the highly efficient, high-gear deeper flexors.