Muscle loads slowly
Slow shortening keeps force high and spreads the work through time.
Life, under a millisecond
The Spring the Muscle Can't Match
Measured froghopper jumps demand tens of thousands of watts per kilogram of jumping muscle, far beyond a conservative 250-500 W/kg comparison band. Change the launch-budget inputs, move body mass through a declared toy motor-spring-latch model, and see which claims survive.
A meadow froghopper is weighed in milligrams. Its jump is not a small version of ours.
12.3 mg
Put the jump on muscle's bill
Malcolm Burrows measured a mean takeoff speed of 2.8 m/s, rising to 4 m/s in the best jumps, with the hind legs extending in less than 1 ms. He measured the paired jumping muscles at about 11% of body mass. Move any of those assumptions below.
Layer 1: the power budget
How much power must each kilogram of jumping muscle deliver?
The body mass stays at Burrows' measured mean, 12.3 mg. The three uncertain or variable terms remain in your hands.
20 µJ140 µJ
0.50 ms1.50 ms
5% body mass20%
Required launch power density
Energy divided by launch time and jumping-muscle mass.
Compared with 250-500 W/kg
The lower ratio uses 500 W/kg. The upper uses 250 W/kg.
P = E / (t × mbody × fmuscle)
The brief's proposed 10-100× range does not survive its own central assumptions. At the measured 11% muscle fraction, Burrows' average and best energy estimates produce a comparison range of about 72-296×. The calculator does not hide that larger result. It also does not turn the 250-500 W/kg band into a universal hard wall. Muscle type, temperature, activation, strain path, and the meaning of peak versus cyclic power all matter.
Power is not energy. A muscle can supply the energy if it is allowed to work for longer. What it cannot do is supply that work directly through the feet in one millisecond. The escape is to move the work earlier.
Move the work to before the jump
The froghopper contracts its trochanteral depressor muscles while the legs remain cocked. A cuticular locking geometry prevents early motion. Elastic structures in the thorax store the work. Release turns a long, low-power loading phase into a short, high-power launch.
The legs wait
A femur-coxa lock holds the cocked leg while elastic energy accumulates.
Release is fast
The stored work leaves on the spring's timescale, decoupled from muscle speed.
Burrows' short 2003 report suggested that the energy could be stored in strengthened thoracic cuticle that might contain resilin, or in the large tendon. Later anatomical work describes the paired pleural arches as important elastic structures. That does not justify reducing every jumping insect to a “resilin spring.” The actual elastic system is composite, three-dimensional, and species-specific.
Make the animal larger
A smart objection is that more muscle should solve the problem, or that springs are simply “an insect trick.” The useful answer is narrower. In a minimal force-velocity model, a small inertial load quickly approaches its motor's unloaded speed. A preloaded spring does not share the motor's speed limit. As the load grows, inertia slows both routes and the spring's peak-power advantage disappears.
Layer 2: a minimal dynamical model
Same geometrically scaled motor, two launch routes
The red curve shows a Hookean spring released from a latch. The blue curve shows the motor loading that spring. The direct motor peak readout gives the same motor's peak power when driving the body directly through the same stroke.
Declared calibration: at the 12.3 mg reference mass, this toy model assigns Burrows' reported 49 µJ average jump energy to the ideal spring. It also repurposes his 34 mN average external launch-force estimate, calculated as body mass times average acceleration, as both the motor's stall force and the Hookean spring's peak force. The 34 mN value is not a measured peak spring force. Giving both routes the same force scale makes their comparison legible, but it is a free modelling choice. It sets the reference stroke to 2E/F = 2.88 mm and the spring stiffness to 11.80 N/m.
0.001 mg100 kg
The slider is logarithmic. The model scales force with length², stroke and unloaded speed with length, and mass with length³.
Spring peak
Direct motor peak
Amplification
Model crossover
Load to 99%
Ideal release
At 12.3 mg, this ideal spring model reaches 70.9 times the direct motor's peak power.
Peak-power amplification across body mass
left: smaller load, larger advantage
horizontal line: equal peak power
Slow loading, fast release
Loading and release power curves update with the chosen mass.
loading uses its own 0-167 ms axis
release uses its own 0-1.60 ms axis
motor: F = Fmax(1 - v/vmax) | spring: m x'' = k(d - x)
Why the crossover is not a zoological border
The live crossover is a consequence of this page's declared scaling rules and calibration. Change spring mass, efficiency, geometry, muscle activation, length-tension effects, gearing, or the performance target, and it moves. Ilton and colleagues used a family of motor, spring, and latch models to establish the general transition, not one universal animal mass.
A later model by Sutton and colleagues asked a different question, jump height rather than peak power, and included a cost this ideal spring omits: muscle length-tension limits on how much work can be loaded. Their model put its crossover below about 2.5 g. The number here is therefore a showing of the mechanism and its dependence on assumptions, not a prediction that every animal below 4.39 kg should latch its jump.
Large animals also use elastic tendons. Deer are not “spring-free.” The narrower observation is that animals above about 1 kg are not documented as launching whole-body jumps primarily from a latched spring. Frogs and many other jumpers can mix muscle and elastic actuation.
The check
The budget and kinematic results below are recomputed in this browser from Burrows' reported inputs. The toy-model outputs follow from the separately stated 49 µJ and repurposed 34 mN calibration plus the minimal equations, not from measured spring or motor parameters. The independent Node verifier repeats the arithmetic and integrates the spring-release ODE from scratch.
Current budget
Kinematic cross-check
Published average
Published best
Calibrated toy ODE
Size crossover
- Measured inputs: Burrows reports body mass 12.3 ± 0.7 mg, mean speed 2.8 ± 0.1 m/s, best speed 4 m/s, less than 1 ms leg extension, energies of 49 and 100 µJ, and jumping-muscle mass near 11% of body mass. These are observations with sampling and rounding, not mathematical constants.
- Muscle comparison band: 250-500 W/kg is deliberately a range. Bird flight measurements reach about 471 W/kg in a maximal burst, and other preparations and definitions give different values. Even the generous edge remains far below the froghopper launch requirement under the central assumptions.
- Power-density inference: dividing external launch power by jumping-muscle mass is a diagnostic test for power amplification. It is not the spring's own material power density and does not mean the muscle produces that instantaneous output.
- Toy-model calibration: the model combines the reported 49 µJ average jump energy with the 34 mN average external launch-force estimate. It repurposes that average force as both motor stall force and ideal spring peak force. Burrows did not measure a 34 mN peak spring force. This choice drives the baseline 48.0 mW spring peak, 70.9× amplification, 1.60 ms release, and 4.39 kg crossover.
- Free model choices: the size model also uses a linear force-velocity motor, a massless ideal Hookean spring, no drag, no gravity during launch, no gearing, perfect latching, perfect energy transfer, 11% muscle at every size, geometric similarity, and a 500 W/kg motor calibration. The toy-model outputs belong to this full set of choices.
- Anatomy: the froghopper uses a cuticular elastic system associated with the pleural arches and a cocking lock. The exact distribution of strain and the contribution of resilin, stiff cuticle, and tendon should not be collapsed into one universal material claim.
Machine check: node research/the-spring-the-muscle-cant-match/verify-the-spring-the-muscle-cant-match.mjs
Sources and honest apparatus
- Malcolm Burrows, “Froghopper insects leap to new heights,” Nature 424 (2003), 509. This is the primary source for the 12.3 mg Philaenus spumarius measurements, the 2.8 and 4 m/s speeds, the less-than-1 ms launch, the 49 and 100 µJ energies, the 11% muscle estimate, and the locking-mechanism interpretation.
- Malcolm Burrows, “Jumping performance of froghopper insects,” Journal of Experimental Biology 209 (2006), 4607-4621. The later five-species study used faster imaging, reported a best Philaenus takeoff of 4.7 m/s in 0.875 ms, and described variation across froghopper species.
- Malcolm Burrows, Stephen R. Shaw, and Gregory P. Sutton, “Resilin and chitinous cuticle form a composite structure for energy storage in jumping by froghopper insects,” BMC Biology 6 (2008), 41. This is the direct anatomical and mechanical evidence for the paired pleural arches as composite elastic stores.
- Mark Ilton and colleagues, “The principles of cascading power limits in small, fast biological and engineered systems,” Science 360 (2018), eaao1082. This supplies the general motor-spring-latch framework, the linear motor force-velocity starting point, and the mass-dependent transition.
- Gregory P. Sutton and colleagues, “Why do Large Animals Never Actuate Their Jumps with Latch-Mediated Springs?” Integrative and Comparative Biology 59 (2019), 1609-1618. This adds muscle length-tension limits and demonstrates why a crossover depends on the performance question and model.
- Sarah J. Longo and colleagues, “Beyond power amplification,” Journal of Experimental Biology 222 (2019), jeb197889. This is the source for the diagnostic logic and its limitations.
Corrections
Corrected 2026-10-14: This page said a small inertial load “quickly outruns its motor.” In the stated direct-drive model, the load approaches the motor's unloaded speed without exceeding it; found by the assay line, 2026-10-14.
Corrected 2026-10-14: This page described the blue curve as direct body drive. It shows the motor loading the spring; direct-drive peak power is shown in the numerical readout; found by the assay line, 2026-10-14.
Corrected 2026-10-14: The 500, 2,500, and 12,500 W/kg ticks were at 25%, 50%, and 75% of the logarithmic scale. They belong at approximately 23.299%, 46.598%, and 69.897%, respectively; found by the assay line, 2026-10-14.