A team of Australian marine biologists were diving off the coast of northern Sulawesi, Indonesia, on a research trip to study a completely different species of octopus, when one of them saw something he could not explain. A small octopus on the seafloor was busy manipulating two halves of a discarded coconut shell. The researcher assumed the animal was going to hide inside one of them. Which would have been ordinary. Octopuses hide inside things constantly. Empty shells, bottles, pipes, anything with a cavity.
What the octopus did instead was stack the two shell halves like nested bowls, position itself on top of them, extend its arms rigidly around the outside, and walk away across the sand carrying the whole assembly with it. The researcher, on his own subsequent testimony, laughed so hard through his regulator that he had to surface.
What the researchers actually documented
According to the 2009 paper by Dr Julian Finn and Dr Mark Norman of Museum Victoria in Melbourne, working with Dr Tom Tregenza of the University of Exeter’s Cornwall Campus, published in Current Biology under the title “Defensive Tool Use in a Coconut-Carrying Octopus”, the researchers spent more than five hundred diver hours between 1999 and 2008 observing veined octopuses, the species Amphioctopus marginatus, on soft-sediment substrates off northern Sulawesi and Bali. Over that period they watched twenty individual octopuses closely. On four separate occasions, individuals were observed carrying stacked coconut shell halves across the seafloor for distances of up to twenty metres.
The carrying method the researchers named stilt-walking. The octopus positions the two shell halves beneath its body with the concave surfaces facing up, extends its eight arms rigidly around and beneath the shells, and walks forward on the stiffened arms. The whole assembly moves together as a single ungainly package. The gait is measurably slower and more energy-costly than the animal’s ordinary locomotion, and it leaves the octopus’s soft body exposed and vulnerable to predators during the entire journey. There is no immediate benefit to the animal while the shells are being carried. Whatever benefit they provide arrives later, once the octopus has stopped moving.
At the destination, the animal inverts the top shell over the bottom one, climbs inside the resulting sphere, and pulls the closure shut around itself, leaving only a small gap for water flow. The two shell halves, now assembled, function as a small hardened chamber inside which the otherwise-defenceless octopus is completely protected from view and from most physical attack. When the animal decides to move again, it disassembles the shelter, re-stacks the halves, and stilt-walks away.
If you enjoy learning about the weird and wonderful parts of our world, check out this video we created on “The Wall Only Elephants Can See”:
The bipedal locomotor ability itself was not new to the scientific literature when Finn’s team published. According to a 2005 paper by Christine L. Huffard, Farnis Boneka and Robert J. Full, published in Science under the title “Underwater Bipedal Locomotion by Octopuses in Disguise”, the same species had been documented four years earlier walking bipedally without any object. Huffard’s team recorded the veined octopus using two of its rear arms in a rolling tank-tread gait, while wrapping the remaining six arms around its body in what looks from a short distance like a small coconut tiptoeing across the sand. The behaviour, on the Berkeley group’s kinematic analysis, was the first case of true bipedal walking ever documented in an animal without a rigid skeleton. What Finn’s team added, on their subsequent field observations, was that the same species could perform a functionally similar coordinated gait while carrying and manipulating external objects, which is the ingredient that pushed the behaviour from unusual locomotion into anticipatory tool use.
The behaviour has been recorded multiple times since the 2009 paper, filmed for BBC Blue Planet II under David Attenborough’s narration, and repeated in laboratory studies of the same species. It is one of the more thoroughly documented behaviours in the entire cephalopod research literature.
Why this counts as tool use, and why the distinction matters
The word tool has been contested in animal-cognition research for as long as anyone has been trying to apply it to non-human species. Otters use rocks to smash open shellfish. Chimpanzees use twigs to fish for termites. Crows bend wire into hooks. New Caledonian crows have been documented manufacturing multi-step tools. Elephants use branches as fly swatters. Dolphins carry sponges on their beaks while foraging.
What has never been reliably documented, until the Finn team’s 2009 paper, is the same class of behaviour in an invertebrate. Which matters because invertebrates diverged evolutionarily from the vertebrate lineage roughly 550 million years ago, in the Cambrian period, before either group had anything like a complex brain. Whatever cognitive capacity underlies the coconut behaviour in this octopus evolved entirely separately from the capacity that underlies tool use in a chimpanzee or a crow. It is not a shared ancestral trait. It is a case of the same cognitive solution arriving at two very different places by two completely independent evolutionary paths.
The specific reason the coconut behaviour has been treated as tool use, rather than as ordinary shelter use of the kind hermit crabs perform every day, sits in the details of how the shells are handled. A hermit crab lives inside its shell continuously. The shell is part of its ongoing life. The veined octopus does something different. It picks up the shell halves at one location. It carries them, at real energetic cost and with no immediate benefit, to somewhere else. It assembles them into a functional shelter only at the point where the shelter is actually needed. Which means the animal is, on the standard definition operating in the cognitive literature, using foresight. It is acquiring an object now for a benefit it anticipates receiving later.
The Finn team’s paper articulated this distinction directly. There is, in their published wording, a fundamental difference between picking up a nearby object and putting it over your head as protection versus collecting, arranging, transporting, and assembling portable armour as required. The second behaviour requires the animal to hold in mind, throughout the carrying phase, a specific future use for an object that is currently causing it only cost.
Which is what tool use, as most cognitive researchers now define it, actually means. It is not merely the manipulation of an external object. It is the anticipatory manipulation of an external object for a future benefit. And the veined octopus, on the field evidence, does this repeatedly and reliably enough that the behaviour is now taught in comparative cognition courses as one of the clearest examples of the phenomenon anywhere in the animal kingdom.
None of which is entirely settled. Some researchers have argued that the coconut behaviour sits closer to a specialised form of shelter construction than to true tool use in the sense in which the term is applied to primate behaviour. The debate is ongoing. What is not really contested any more is that a small soft-bodied mollusc, in the shallow tropical waters of Indonesia, has been observed carrying out a sequence of coordinated actions that no one had previously believed a wild invertebrate was capable of performing.
The animal weighs a few hundred grams. Its brain contains around five hundred million neurons, which is roughly the same number as a domestic cat, though the neurons themselves are distributed very differently, with about two-thirds of them located in the arms rather than in a central brain. Whatever cognitive machinery is producing the coconut behaviour is running on hardware that is not vertebrate, is not centralised, and is not closely related to any of the other animals now known to use tools. Which is why the finding mattered when it was first published. And why the small pink-veined octopus in the videos, stilt-walking awkwardly across the sand carrying its little portable house, continues to disturb the previously comfortable assumption that this kind of behaviour belonged only to warm-blooded animals with backbones.
Kiran Athar is a writer, not a marine biologist or a cognitive scientist. This piece draws on primary peer-reviewed research in Current Biology and Science.
