A bottlenose dolphin in open ocean detects a 7-centimeter ball bearing at 113 meters. In the same conditions, it distinguishes between a brass ball and an aluminum ball of identical size by the difference in their echo. These are not laboratory curiosities — they are measurements from actual behavioral experiments, and they place dolphin echolocation somewhere above what any human-made active sonar system can do at comparable range and target size.
The engineering inside a dolphin's head is worth understanding on its own terms.
The Signal
Dolphins produce clicks rather than sustained signals. Each click is a broadband pulse, typically in the range of 40 to 130 kHz, lasting between 50 and 200 microseconds. The peak energy is well into ultrasound — human hearing cuts off around 20 kHz. The click rate varies with context: in open water searching, dolphins emit clicks at around 200 per second. As they close in on a target, the rate drops to allow each echo to return before the next click is sent. During the final approach to prey, it can drop to 10 or 20 clicks per second with the animal clearly tracking something specific.
The clicks are produced not in the larynx — dolphins have no vocal cords — but in the nasal passage, using a pair of phonic lips (also called monkey lips or MLDB: museau de singe) located just inside the blowhole. These are fatty-tissue structures that vibrate when air is pushed across them from air sacs behind. The dolphin doesn't exhale to produce clicks; the same air is recycled back and forth. A dolphin can echolocate continuously without breathing, using internally circulating air.
The Melon
The forehead of a dolphin — the melon — is an acoustic lens. It is a mass of lipid tissue with a specific density gradient: less dense at the center, denser toward the edges. This gradient bends sound in a way analogous to a glass lens bending light. Clicks produced at the phonic lips are shaped by the melon into a forward-directed beam.
The beam is not diffuse. Measurements using hydrophone arrays show that the dolphin's transmitted click is concentrated into a cone roughly 10 to 20 degrees wide at peak energy. This directionality serves two purposes: it concentrates energy toward the target, increasing detection range, and it reduces acoustic clutter from reflections outside the beam. The dolphin is pointing a flashlight, not turning on a room light.
The melon shape is under muscular control. Dolphins can, to a limited degree, steer the beam by changing melon shape and head orientation. The steering range is not large — they primarily use body orientation to aim — but the ability to modify beam shape is real and adds a degree of flexibility that fixed-geometry sonar hardware doesn't have.
Receiving the Echo
The dolphin's jaw is hollow. The lower jaw contains a fat-filled channel that runs from the tip of the rostrum back toward the ear. Returning echoes are received through this channel — not primarily through the external ear canal, which is narrow and may not be the dominant acoustic path. The fatty channel conducts sound efficiently from the jaw tip to the inner ear.
The inner ear itself is housed in a dense bone (periotic bone) that is acoustically isolated from the skull. In most mammals, sound reaches the cochlea partly through bone conduction from the skull. In odontocetes (toothed whales and dolphins), the ear is acoustically decoupled from the skull by sinuses filled with foam and fluid, which reduces acoustic cross-contamination between the two ears. This isolation is what gives dolphins fine-grained directional hearing — they can localize a sound in three dimensions because the two ears receive clean, separate signals.
Processing the Return
The auditory system of dolphins processes temporal information with extremely fine resolution. The inter-aural time difference — the delay between the echo arriving at the near ear versus the far ear — can be as small as a few microseconds. Dolphins can resolve these differences, which is what allows them to determine the angle to a target with precision measured in fractions of a degree.
More interesting is the frequency-dependent information in the echo. Different materials have different acoustic impedances, which means they reflect sound with different phase and frequency characteristics. Bone, gas, and water all reflect differently. A fish's swim bladder — a gas-filled organ — has very high acoustic impedance contrast with the surrounding tissue and returns a strong, characteristic echo. Dolphins appear to use this to distinguish species of fish even when size is identical. Some experimental work suggests they can detect internal structure — bone arrangement, organ position — in the same way that medical ultrasound images internal anatomy.
The Cognitive Piece
None of the acoustic hardware is useful without the brain processing to interpret it. Odontocete auditory cortex is large and anatomically distinct from terrestrial mammal auditory cortex. The inferior colliculus — a key relay station in the auditory pathway — is disproportionately large in dolphins compared to other mammals of similar body size. This suggests substantial central processing dedicated to acoustic analysis.
The processing is fast. A dolphin tracking a moving fish in turbid water is updating its acoustic picture 20 or more times per second, computing range and angle from successive echoes, and translating that into motor commands for pursuit. The reaction time from echo detection to course change is measured in milliseconds. This is not just detection — it is three-dimensional tracking with continuous motion prediction.
Training experiments by William Au at the University of Hawaii over several decades have mapped the boundaries of what this system can do. The results consistently show performance above what sonar engineers expected from biological tissue — particularly in terms of range discrimination (detecting objects at slightly different distances) and material discrimination. The exact computational mechanisms behind these capabilities are still not fully understood.
Convergent Invention
Echolocation evolved independently in bats (order Chiroptera) and in odontocetes at least 35 million years ago. The two solutions share the same acoustic principles — broadband pulses, focused transmission, specialized reception — but the hardware is entirely different. Bats produce clicks in the larynx. Dolphins produce them in the nasal passage. Bats receive through the outer ear. Dolphins receive through the jaw. The similarity is functional, not structural.
The independent convergence on the same acoustic strategy in two entirely different lineages suggests something important: this particular engineering approach — short broadband pulses, focused transmission, high-temporal-resolution reception — is a near-optimal solution to the problem of locating objects in darkness using sound. Evolution found it twice. The solution space for biosonar appears to be narrowly constrained.
We have built active sonar systems that are louder, that operate at longer range, and that carry more processing power. We have not built systems that match dolphin echolocation for resolution, material discrimination, or energy efficiency in a comparable form factor. The engineering is better than ours, still, fifty million years in.
Curious about what it looks like to build something in public? We publish our milestones, limitations, and progress at builds.anethoth.com.