Marius Comper

The physics inside a familiar gesture

A Wet Mouse Shakes 30 Times a Second

A Labrador needs about 4.5 Hz and a brown bear 4 Hz. The adult mouse measured 29 Hz, rounded to 30 in the title; the physical model explains the rhythm through the acceleration needed to release water from fur.

Dickerson, Mills and Hu, 2012 · 33 animals · 16 species · filmed at 500–1,000 frames per second

Chronophotography plate · one second

Labrador 2 in motion

Mass 28.1 kg
Frequency 4.5 Hz
Acceleration 15 × g
Cycle length 222 ms
Body oscillations within one second The wave becomes denser for small animals and sparser for large animals. 0 s 1 s

Each yellow band represents one complete cycle in the measured second. A paler final band shows the remaining fraction. Frequencies and accelerations come from Table 1 of the study.

Size sets the rhythm

The sample runs from 0.01 kg to 260 kg, a ratio of 26,000 to 1 and more than four orders of magnitude. Frequency falls predictably across that enormous span.

fM−0.22 fit across 25 measured points · R² = 0.95

The observed exponent, −0.22, sits close to −3/16, or −0.19 when rounded, which the authors derive from the balance between drop inertia and surface tension. The relationship describes a strong trend, while each animal's anatomy accounts for deviations around it.

The adult mouse in the table shakes about 6.4 times as fast as Labrador 2. One cycle lasts 34 ms in the mouse and 222 ms in the Labrador.

The drop leaves when inertia beats capillarity

Water wets hair and gathers between fibres. Surface tension holds each drop like an elastic film. A rapid change of direction gives the drop an outward apparent force.

Tension holds
Capillary force pins water between clumps of hair.
Inertia releases
The acceleration 2 increases the drop's apparent weight.

Values in the table range from 12 to 72 times gravitational acceleration. The authors summarise the useful domain as roughly 10–70 × g, surprisingly narrow for such different animals.

If the adult mouse copied a dog's rhythm and stopped at 4 Hz, the calculation gives about 1 × g. Water would remain attached to the fur.

Loose skin turns the torso into a whip

A Labrador keeps its paws on the ground while skin slides far beyond the skeleton's movement. A drop at the end of a hair therefore receives speed that the spine's rotation alone could not produce.

90°amplitude of the marker on the skin
30°estimated amplitude of the spine
gain in speed-related force

Skin motion reaches about three times the amplitude of vertebral movement. The force associated with speed grows with the square of amplitude, giving the factor of nine.

One second of effort saves a great deal of heat

Wet fur can retain a mass of water that is expensive to evaporate. Shaking ejects the mechanically removable part and leaves a thin film for evaporation.

70%of accumulated water is removed
30%remains in the fur

The authors' energy model estimates that shaking uses between one ten-thousandth and one thousandth of the energy required to evaporate the removable water. In percentage terms, that is 0.01%–0.1%.

The same outcome through very different rhythms

The goat, pig and sheep sit far from neighbours of similar mass. Body shape, fur and shaking style matter, while the scaling relationship remains clear across all 25 points.

AnimalMassRateAcceleration
Adult mouse0.0272 kg29 Hz66 × g
Cat3.3 kg9.4 Hz33 × g
Labrador 228.1 kg4.5 Hz15 × g
Boer goat48.3 kg7.7 Hz50 × g
Kunekune pig49.4 kg8.2 Hz57 × g
Black bear90 kg4.1 Hz16 × g
Brown bear260 kg4 Hz24 × g

A strong fit still has blurred edges

Most species are represented by a single animal. The squirrel's and both bears' mass and radius were estimated from published literature; the lion's and tiger's radius was estimated from similarly sized adults.

Very large animals gradually depart from the scaling relationship. They retain heat better, have different skin layers and may carry less hair. The authors also observed that hairless guinea pigs shivered instead of performing the full shake.

The study explains the mechanics of drying after the gesture begins. The sensory circuit that triggers the movement lies outside this demonstration.

Fast cameras, fur markers and a simulator with wet brushes

The researchers filmed animals immediately after wetting, tracked markers fixed to fur and measured mass, torso radius, frequency and acceleration. A laboratory device spun fur samples and wet brushes to test when drops detached.

The relationship between mass and frequency was fitted with one representative point for each species or breed, apart from the separately measured Labradors. Comparisons on this page use the published values, with rounding repeated after calculation.

Source accessed 29 August 2026.

  1. Andrew K. Dickerson, Zachary G. Mills and David L. Hu, “Wet mammals shake at tuned frequencies to dry”, Journal of the Royal Society Interface, 2012.
  2. PubMed record and abstract.