Sea Otters Use Rocks as Tools, What That Reveals About Mammalian Intelligence and Cognition

Before a sea otter dives for a clam, it often tucks a flat stone under its forearm, a pocket it keeps specifically for this purpose. On the surface, it places the stone on its chest, holds the shellfish with both paws, and hammers until the shell cracks. Then it eats, stashes the rock again, and dives for the next one. The same rock. The same otter. Sometimes for days in a row.

This is not a fluke. A 2017 study published in Biology Letters, led by primatologist Michael Haslam and his team at the University of Oxford, documented stone tool use in wild sea otters at Elkhorn Slough in California. The researchers found that individual otters showed strong preferences for specific rocks, selecting them by size, weight, and surface texture relative to the prey they were targeting. Otters cracking hard-shelled mussels chose heavier stones than those working on softer clams. That is not random behaviour. That is selection based on a physical problem.

Biologists define tool use precisely: an animal must externally deploy an unattached object to alter the state of another object, and the user must hold or carry the tool, not simply use a fixed surface. Sea otters (Enhydra lutris) meet that definition. They are one of fewer than ten non-primate mammal species confirmed to use tools in the wild, a list that also includes bottlenose dolphins, which carry marine sponges on their snouts to protect against the seafloor while foraging, and a handful of corvid birds, which are not mammals at all.

The otter's anatomy makes the behaviour more striking. Unlike primates, otters have no opposable thumbs. Their forelimbs are short and their grip is built for swimming, not manipulation. The fact that they have adapted tool use around this physical constraint points to something cognitive compensating for what the body cannot do automatically.

Pups do not hatch knowing how to use rocks. Research tracking mother-pup pairs shows that young otters learn the behaviour by watching their mothers, then practising with objects, often poorly at first, dropping rocks, misjudging angles, losing prey. The learning curve is real and takes months.

This is called social transmission, and it is the same mechanism by which human children learn to use cutlery, or by which chimpanzees in West Africa learn to crack nuts with stone hammers. The presence of social transmission in a species is significant because it means the behaviour is cultural, not purely genetic. If an isolated otter raised without a mother who used tools were placed in a shellfish-rich environment, there is no guarantee it would develop the behaviour independently. The knowledge travels between individuals, not just between generations through DNA.

A 2020 paper in PLOS ONE by Fragaszy and colleagues, examining tool traditions across mammals, placed sea otter rock use in the same category as chimpanzee nut-cracking: a socially maintained technical tradition. That framing matters. It means the rock-and-shellfish system is not a fixed species trait, it is a local practice that could, in theory, disappear from a population if the transmitting individuals were removed.

The sea otter case complicates a tidy story about intelligence. For decades, tool use was treated as a marker of primate exceptionalism, something that separated apes and humans from the rest of the animal kingdom. Otters are not primates. Their brains are structurally different. Their evolutionary lineage diverged from ours far earlier. Yet they solve the same class of physical problem, hard shell, soft food inside, need an intermediary, using the same general strategy primates use.

This is called convergent cognition: independent evolutionary lines arriving at similar cognitive solutions because the environmental pressures are similar. The otter did not inherit tool use from a common ancestor with chimpanzees. It arrived there separately. That convergence is what makes the behaviour scientifically interesting. It suggests that certain forms of problem-solving intelligence are not the exclusive property of any one lineage. They emerge wherever the ecological conditions make them worth developing.

Otters also show individual variation in skill level, some are faster, some choose better rocks, some crack shells with fewer strikes. That variation is the raw material on which learning and selection operate. A population of otters is not running identical software. Each animal is working something out for itself within a tradition it inherited socially.

The study of tool use in animals has shifted considerably since Jane Goodall first documented chimpanzee tool use at Gombe in 1960. Researchers now track tool behaviour in crows, octopuses, elephants, and otters, species whose brain architectures look nothing like a primate's. Each new confirmed case weakens the assumption that complex cognition requires a specific brain structure or a minimum brain size relative to body weight.

Sea otters add a particular wrinkle. They are marine mammals that returned to the ocean from a terrestrial ancestor. Their forelimbs, which a land-dwelling ancestor used for walking and grasping, became flippers optimised for swimming. Tool use, in this context, is the animal finding a workaround for a body that evolution repurposed. The cognition is filling a gap the anatomy left open.

What the rock in the armpit finally reveals is that intelligence in mammals is less a fixed endowment than a set of responses to specific pressures, social, physical, ecological. The otter did not develop tool use because it is smart in some general sense. It developed tool use because it lives where shellfish are abundant, where shells are hard, where mothers can be watched, and where carrying a rock costs less than going hungry. Strip any one of those conditions away, and the behaviour may never appear. Put them all together, and even an animal with flippers works out how to use a hammer.