What Adam Is Reading
How Does an Octopus Fit into a Beer Can?
Three French octopuses, four Mexican ones, a Plexiglas wall with a hole in it, and a Singapore robotics lab that wants to know how they do it. The answer is one arm at a time. The bigger questions remain open.
Things I Learned This Week· Journal of Experimental Biology, Short Communication· 6 sources· November 2026

An octopus has no skeleton. The only hard part is the beak, and a cartilage capsule around the brain. Divers have filmed them climbing into beer bottles and soda cans for years, and the videos do well. Nobody had written down how they actually do it.

Now somebody has. A team from the National University of Singapore, the University of Naples and a Vienna aquarium put octopuses on one side of a clear acrylic partition with a single hole in it and filmed what happened from two angles at 30 frames per second. The paper is five pages long. It has a Fisher exact test in it. It is the first formal description of octopus squeezing.

The study at a glance
Who3 subadult common octopuses (Octopus vulgaris), caught by fishermen in La Ciotat, France
Also4 juvenile Mexican four eyed octopuses (Octopus maya), 4 months old, raised from eggs
WhereHaus des Meeres, Vienna, a public aquarium inside a World War II flak tower
Trials33 recorded for O. vulgaris, 30 usable
Size measureDistance between the eyes, a stand in for the brain capsule
FundingAn NUS start up grant and a Bridging fund

What they found

The octopuses used one of two routines, and which one depended on whether the hole was wider than the distance between their eyes.

  1. Probe. One arm goes through first, bending into a U at the rim and then reaching out the far side. Usually a front arm, and in this tank usually the second arm on the left (47 percent of trials).
  2. Shift. Three to five more arms follow (four was most common) until the hole is essentially plugged with octopus.
  3. Squeeze. The arms on the far side twist back, stick their suckers to the partition, and stiffen. The octopus pulls itself through. Representative frames show the whole pass taking about five seconds.
  4. Squeeze again (small holes only). When the hole is narrower than the eyes, step three splits in two. The animal rotates its body to wedge the inserted arms deeper and widen the gap, works the remaining arms through, then rights itself and hauls the head and mantle through last. About twenty seconds.

Side views show the mantle actively shrinking its own diameter on the way through. That is muscle at work, on top of the squish. Circular muscles in the mantle and transverse muscles in the arms let the animal choose its cross section.

The Mexican octopuses were a different story. Offered the small hole, they declined. They tried a little, retreated to a corner of the box and stayed there. Offered a hole of each size at once, one tried the small hole first, got two arms through, failed, and moved over to the big one, where it ran the three step routine without trouble.

The finding worth keeping. Folk wisdom (and the octopus pot fishery literature) says that if the beak fits, the octopus fits. The authors suggest the real limit may be the combined cross section of all eight arm bases where they meet the body. They also note the animals do not seem to judge in advance whether a hole is passable. The octopus overestimates itself and corrects in the field. Most of us run the same algorithm in a parking garage.


How much weight it bears
1
Hole size determines the strategy (P < 0.001)
What the paper shows

In the larger hole condition, 14 of 14 trials used three steps. In the smaller hole condition, 15 of 16 used four. That is a clean split.

The gap

The larger hole condition was one octopus. The smaller hole condition was the other two. Each animal was tested in only one condition, so the statistics compare animal 1 with animals 2 and 3. The authors say so plainly and cite the 3Rs (reduce the number of animals used) as the reason. Hole size and individual personality are fully tangled.

Mostly Solid
2
"Larger" and "smaller" are relative
What the paper shows

The larger hole was 20 mm across, for an octopus whose eyes sat 15 to 20 mm apart. The smaller hole was 30 mm, for octopuses whose eyes sat 30 to 35 mm apart. The labels describe the fit, not the hole.

The gap

The "smaller hole" was half again as wide as the "larger hole." This is correct and sensible. It is also the kind of sentence that makes a methods section read like a riddle.

Solid
3
Species comparison
What the paper shows

The four step strategy appeared only in O. vulgaris. O. maya passed only through relatively large holes.

The gap

Species, age and upbringing all changed at once. The French animals were wild caught adolescents. The Mexican animals were four month olds hatched and raised in a tank in Vienna, with no prior experience of rocks, crevices or recycling bins. The paper presents the O. maya trials as exploratory, which is the right label.

Mostly Solid
4
The things the authors noticed but could not test
What the paper shows

Three of the 33 trials failed because the octopuses damaged the apparatus trying to escape somewhere other than the hole. An animal that escaped by squeezing out of the container entirely was scored as a negative. The strong left arm preference may reflect the tank itself, since one side offered a clearer view out. And in the four step condition, passes got slower over successive trials. The authors offer growth or "declining exploratory motivation" as possible explanations.

The gap

Translated, the subjects broke the equipment, ignored the assigned exit, favored the side with the view, and got bored. Anyone who has run a clinical trial will recognize the cohort.

Solid

Why the engineers care

Search and rescue robots have a bottleneck problem, literally. A rigid robot has to be built small enough for the narrowest gap it will ever meet, and shrinking it costs sensors, grip and power. So rescue robots trend toward snakes.

The octopus solves the problem by serializing. It sends one part through, anchors it, and uses it to pull the rest. The paper argues this looks like a higher level command that strings together simpler motor patterns, a control scheme that might let a soft robot with many floppy limbs coordinate without computing every joint at once. The senior author, Cecilia Laschi, built one of the first eight armed soft octopus robots a decade ago.


The questions the paper leaves open

The study establishes how. It is silent on why, where and which brand. Some preliminary work exists.

Why. An octopus is a soft, edible animal in a world full of things with teeth. It wants a den with a narrow door. A beer can is a den with a narrow door that someone has helpfully left on the seafloor. Fishermen understood this long before roboticists did. Clay octopus pots on a line, the Japanese takotsubo and its Mediterranean cousins, work because the animal cannot resist a hollow. One of the references in this paper is a fisheries study of which pot designs catch which size of giant Pacific octopus.

Where. A 2022 study in Marine Pollution Bulletin gathered 261 underwater images and videos of octopuses interacting with litter and found 24 species using it, most often as shelter. Glass led at about 40 percent, plastic followed at about 25, and cans, broken bottles and old batteries all made appearances. The coconut octopus was the most frequent tenant. The ocean has a housing market, and we have been supplying the inventory.

Which beer. Here the literature is thin. The study's own sample suggests a design. The French octopuses came from La Ciotat, where pétanque was invented in 1907 and the Lumière brothers filmed a train arriving in 1895. The Mexican octopuses trace back to Sisal on the Yucatán coast. They were housed in Vienna and analyzed by engineers in Singapore and a biologist in Naples. One can per nation, hole diameter scaled to eye spacing, would settle the cultural question in an afternoon. I note only that the French animals tackled the harder opening and the Mexican animals held out for something roomier, which may say more about can design than character.

Who pays. The work was funded by a university start up grant and something called a Bridging fund, which is the right name for money that gets you through a narrow gap. The authors declare no competing interests. No brewery is listed.

So What

An octopus fits into a beer can the way most hard problems get solved. One part goes through first, finds a grip, and pulls the rest after it.

Confidence: the motor sequence is well documented on video and plausible across species. The claim that hole size selects the strategy rests on three animals, one per condition in the key comparison, and the authors say so. The beer preference material is the author's own extrapolation and has not been peer reviewed.

Sources

Primary paper: Xin W, Kuba MJ, Gutnick T, Abed-Navandi D, Laschi C. Overcoming environmental constraints: octopus uses 3-step or 4-step strategies to squeeze through small holes. J Exp Biol 229 (2026) jeb252825. doi:10.1242/jeb.252825

Coverage: Reese D. How an Octopus Can Squeeze into a Beer Can. Nautilus, 6 Nov 2026. nautil.us

Octopuses and marine litter: Proietti M and colleagues, Federal University of Rio Grande. Marine Pollution Bulletin 2022, as reported in Smithsonian. smithsonianmag.com

Soft octopus robot: Cianchetti M, Calisti M, Margheri L, Kuba M, Laschi C. Bioinspired locomotion and grasping in water: the soft eight-arm OCTOPUS robot. Bioinspir Biomim 10 (2015) 035003. doi:10.1088/1748-3190/10/3/035003

The aquarium: Haus des Meeres. Wikipedia

La Ciotat: pétanque (1907) and the Lumière film (1895). Wikipedia