How Sharks Sense Electricity
A shark can find a fish buried in sand, in the dark, by the faint electricity leaking out of its body. It is the most sensitive electrical sense known in any animal.
A sense we do not have
Junior level — plain language, no maths
Everything alive leaks a little electricity. Your muscles run on tiny electrical pulses, your heart fires one every beat, and even sitting perfectly still your body's chemistry produces a faint electrical field in the water around you. On land this leaks away into nothing. In salt water - which conducts electricity beautifully - it spreads out into the sea like a smell.
Sharks read it. Scattered across a shark's snout are hundreds of tiny pores, each one the opening of a jelly-filled tube called an ampulla of Lorenzini. They were first drawn in 1678 and nobody worked out what they were for until the 1960s. They are voltmeters - and they are extraordinary ones, sensitive to about five billionths of a volt across a centimetre. That is like detecting a 1.5 V battery with one terminal dipped in the sea off Florida and the other off Ireland.
What this buys a shark is the ability to find food it cannot see, smell or hear. A flatfish lying completely buried under sand, holding still, breathing slowly, is invisible and silent - but it cannot switch off its own body. In the simulation above, bury the fish deeper and watch the signal collapse: the field weakens as the cube of distance, so doubling the distance divides the signal by eight. That brutal falloff is why the electric sense only works in the last half-metre. It is not a long-range detector; it is the final aim.
And it explains one of the strangest-looking animals in the sea. A hammerhead's head is not for hammering. Voltage is a difference between two points, so the further apart your two sensors are, the bigger the reading. Spreading the pores across a metre-wide head gives a hammerhead a far longer measuring stick than an ordinary shark - and a bigger reading from the same faint field. Slide the head width up in the simulation and watch a signal that was invisible come out of the noise.
Things worth knowing
- The pores on a shark's snout were drawn by Stefano Lorenzini in 1678. It took nearly three hundred years for anyone to work out that they detect electricity.
- Sharks respond to fields as weak as 5 nanovolts per centimetre — roughly the field of a 1.5 V battery with its terminals dipped in the ocean hundreds of kilometres apart.
- A hammerhead's head is a wider measuring stick. The same faint field produces a bigger voltage across a metre-wide head than across a narrow one.
Ampullae, dipole fields and the inverse cube
Student level — the core equations
An ampulla of Lorenzini is a canal filled with a highly conductive glycoprotein gel, running from a surface pore to a bulb of receptor cells embedded deep in the head. The gel makes the canal an almost perfect conductor, so the receptor at the bottom sits at the potential of the seawater at the pore. Each ampulla therefore measures the voltage difference between its own pore and the shark's internal reference - and because the canals run in many directions and to many different lengths, the array samples the field over the whole head.
Prey generates a roughly dipole field: ion flux across gills and mucous membranes produces standing DC potentials of the order of hundreds of microvolts at the body surface, modulated by ventilation. A dipole field falls as \(1/r^3\). Take a typical prey field of about \(100\ \mu\text{V/cm}\) at 1 cm; at 10 cm that is \(0.1\ \mu\text{V/cm}\), and at 25 cm it is already down to about \(6\ \text{nV/cm}\) - right at the behavioural threshold Adrianus Kalmijn measured in the 1970s and 1980s. The predicted detection range of a few tens of centimetres is exactly what sharks show in feeding trials, where they will strike buried electrodes in preference to real, odorous food.
The voltage an animal actually resolves is \(\Delta V = E \cdot d\), the field multiplied by the separation of the sampling points. This is the quantitative core of the cephalofoil hypothesis for hammerheads: widening the head from 30 cm to 100 cm multiplies the available signal by more than three for the same field. Since the field falls as \(r^{-3}\), a factor of three in sensitivity buys only \(3^{1/3} \approx 1.5\) in range - a real but not spectacular gain, which is part of why the hypothesis remains debated alongside enhanced binocular vision and manoeuvrability.
The same organ has a second job. Moving a conductor through a magnetic field induces a voltage: a shark swimming at 1 m/s through the Earth's 50 µT field generates about \(0.05\ \mu\text{V/cm}\), well above threshold. Elasmobranchs may therefore read a magnetic compass through the same electroreceptors, by induction rather than by any dedicated magnetic organ.
Key Formulas
| Behavioural threshold | \(E_{\min} \approx 5\ \text{nV/cm}\) | Kalmijn |
|---|---|---|
| Dipole falloff | \(E(r) \propto \dfrac{1}{r^3}\) | |
| Prey field | \(E(1\,\text{cm}) \sim 100\ \mu\text{V/cm}\) | |
| Voltage across the head | \(\Delta V = E \cdot d\) | d = pore separation |
| Range gain from width | \(r \propto d^{1/3}\) | the cephalofoil argument |
| Motional induction | \(E = v B\) | 1 m/s in 50 µT gives 0.05 µV/cm |
Things worth knowing
- A dipole field falls as 1/r³, so doubling the distance divides the signal by eight. That is why electroreception only works in the last few tens of centimetres — it is a terminal-aim sense, not a search sense.
- In Kalmijn's experiments sharks ignored real, odorous food and struck at buried electrodes reproducing a prey's electric field. The electric cue overrode smell entirely.
- Swimming at 1 m/s through the Earth's field induces about 0.05 µV/cm across a shark — ten times its threshold. Sharks may read a magnetic compass with the same organ they hunt with.
Ampullary transduction, noise limits and the physics of an impossible measurement
Scholar level — full mathematical depth
01The organ
Each ampulla is a canal of length up to 20 cm in large sharks, filled with a keratan-sulphate proteoglycan gel of resistivity close to that of seawater - about \(25\ \Omega\,\text{cm}\) - and sheathed in an epithelium of extremely high resistance. The design is a deliberate voltage divider: the canal presents negligible series resistance while the wall presents enormous shunt resistance, so essentially the entire external potential difference between pore and body interior appears across the sensory epithelium at the ampullary bulb. The receptor cells are secondary sensory cells with voltage-gated calcium channels and a calcium-activated potassium conductance, arranged so that the membrane sits on a steep region of its \(I\)–\(V\) curve.
02Sensitivity against thermal noise
The claimed sensitivity strains credulity until you do the noise accounting. Johnson–Nyquist noise across a resistance \(R\) in bandwidth \(\Delta f\) is \(V_n = \sqrt{4k_BTR\,\Delta f}\). Ampullary responses are tuned to low frequencies - the useful band is roughly 0.1 to 10 Hz, matching prey ventilation - and the effective source resistance is low. Restricting \(\Delta f\) to a few hertz and averaging across hundreds of ampullae in parallel brings the noise floor below the signal. The sense is not achieved by a miraculous single receptor but by a narrow bandwidth, a low-noise geometry and massive parallel redundancy.
03Why DC coupling is the hard part
Prey fields are essentially DC, modulated slowly. Any real amplifier drifts, and the animal's own muscle and gill potentials are far larger than the signal of interest. The ampullary system solves this with a common-mode rejection architecture: the shared internal reference means a uniform field imposed on the whole animal produces no differential response, while a local source produces different potentials at different pores. It is a differential amplifier built from geometry, and it is why the *spatial gradient* rather than the absolute potential is the operative stimulus.
04The cephalofoil, quantitatively
The electrosensory hypothesis for hammerhead head shape gives a testable scaling. Signal goes as \(\Delta V = E d\), so widening \(d\) from 0.3 m to 1.0 m raises signal 3.3-fold; against an \(r^{-3}\) field this converts to a range gain of \(3.3^{1/3} = 1.49\), roughly 50%. Additionally the swept search area per unit distance travelled scales linearly with \(d\). Kajiura and Holland found hammerhead sensitivity per ampulla no better than in carcharhinids, supporting a geometric rather than a physiological advantage - though the competing explanations of enhanced binocular overlap and improved pitching manoeuvrability remain live, and the shape is probably overdetermined.
05Induction-based magnetoreception
Motional induction gives \(\vec E = \vec v \times \vec B\) in the animal's frame. At \(v = 1\ \text{m/s}\) in \(B = 50\ \mu\text{T}\) the induced field is \(50\ \mu\text{V/m} = 0.05\ \mu\text{V/cm}\), an order of magnitude above threshold. An elasmobranch executing turns can in principle extract heading from the modulation. The persistent difficulty is that the animal must distinguish its own induced field from ocean currents moving conducting seawater through the same field, which generate comparable potentials - and that ambiguity, rather than sensitivity, is the open question.
06The costs of a sense this good
Extreme electrosensitivity has consequences the animal did not sign up for. Submarine power cables, ship hulls with galvanic potentials and the corrosion currents of steel structures all produce fields far above threshold; sharks investigate and sometimes bite them. The same physics underlies electropositive metal deterrents trialled as bycatch mitigation on longlines. A sense evolved to find a flatfish under 20 cm of sand cannot help but also detect an industrial seascape it never evolved with.
Key Formulas
| Threshold field | \(E_{\min} \approx 5\ \text{nV/cm} = 0.5\ \mu\text{V/m}\) | |
|---|---|---|
| Dipole field | \(E(r) = \dfrac{p}{4\pi\sigma r^3}\,\sqrt{1+3\cos^2\theta}\) | |
| Signal across the head | \(\Delta V = E\,d\) | |
| Thermal noise floor | \(V_n = \sqrt{4k_B T R\,\Delta f}\) | narrow Δf is the trick |
| Canal gel resistivity | \(\rho \approx 25\ \Omega\,\text{cm}\) | as conductive as seawater |
| Motional induction | \(\vec E = \vec v \times \vec B\) | |
| Cephalofoil range gain | \(r \propto d^{1/3}\) | 3.3× width → 1.5× range |
Things worth knowing
- The sensitivity works because of narrow bandwidth. Ampullae are tuned to roughly 0.1–10 Hz — prey ventilation rates — and restricting bandwidth is what pushes Johnson–Nyquist thermal noise below the signal.
- The array is a differential amplifier built from geometry: a shared internal reference means a uniform field produces no response, so the animal reads spatial gradients rather than absolute potentials.
- Sharks bite submarine cables and steel structures because galvanic and corrosion currents produce fields orders of magnitude above threshold. The sense cannot tell industry from prey.