Why Do Cats Like Boxes? Some Sit in Squares Nobody Drew

Four pet cats have been filmed sitting inside a square that was not there. Their owners had taped four black Pac-Man shapes to the floor with the cut-out corners facing inward, the arrangement that makes human eyes see a square whose sides are never drawn. In daily trials between June and August 2020, those four cats sat inside its missing edges seven times.

A real square outline drew cats eight times. The experiment, published in 2021 by Gabriella Smith of Hunter College in New York and two colleagues, used the pull of boxes, laundry baskets and taped outlines to test what cats see.

Why do cats like boxes?

Cats like boxes as places to hide, and hiding lowers a cat’s stress. In two Dutch shelter trials, newly arrived cats given a hiding box settled days sooner than cats without one. That is the best-tested reason, though only in shelter cats in their first two weeks. Warmth and a snug fit are the other usual answers.

Smith’s team, writing about the pull of enclosed spaces, was plain about how much is known: “The reason for this behavior is still unknown but is clearly highly desirable.”

Shelter cats with a box calmed down days sooner

The strongest evidence comes from shelters in the first weeks after a cat arrives, when, in the words of a 2014 Utrecht University study, “novelty stress is highest.” Claudia Vinke and two colleagues gave a hiding box to 10 of 19 newly arrived cats at a Dutch shelter and scored them on days 1 to 5 and again on day 14, using the Kessler and Turner Cat-Stress-Score, a scale running from 1, fully relaxed, to 7, terrorized.

The cats with boxes had lower stress scores on days 3 and 4. A follow-up at a shelter in Arnhem, run in November and December 2015 and published in PLOS ONE in 2019, gave 12 of 23 cats a cardboard box 44 by 31 by 26 centimeters with two entrances. The box cats reached a steady stress level seven days earlier than the rest, in the authors’ summary.

	2014 trial	2019 trial
Cats, with box / without	10 / 9	12 / 11
Stress scored on	Days 1-5 and 14	Days 1, 2, 3, 5, 7, 9 and 12
Stress with a box	Lower on days 3 and 4	Lower on every scored day after day 1
Box group settled	Around day 3; groups level by day 14	Day 2, against day 9 without a box

The box did not fix everything. In the 2019 trial nearly all the cats lost weight in quarantine, 6.3% of their starting weight with a box and 7.7% without, and the box made no significant difference to weight loss, adoption rates or length of stay.

Cats with no box improvise. The 2019 paper notes that stressed cats housed with nowhere to hide show more alert resting behind the litter box, which it interprets as alternative hiding behavior.

The snug-fit theory rests on kittens and pigs

Nicholas Dodman, a professor emeritus at Tufts University’s veterinary school, set out the snug-fit explanation in The Conversation in April 2017. He looks back to kittenhood, when cats snuggle against their mother and littermates for warmth and soothing contact, and compares the habit to swaddling. Kneading, the paw-pumping called making biscuits, starts in the same place, as nursing kittens push for milk.

His proposed mechanism is chemical: “The close contact with the box’s interior, we believe, releases endorphins.” The support he gives for that comes from pigs. He wrote that he and Temple Grandin found naltrexone, a drug that counteracts endorphins, “reversed the soporific effect of gentle squeezing of pigs.”

Their 1989 paper is more modest. Ten piglets given naltrexone took about twice as long to relax in a padded squeeze chute as ten given saline, 312 seconds against 162, but ended up just as relaxed, and the authors wrote that the drug “partially blocked” the response. Like the snake theory behind the cucumber videos, the idea borrows its evidence from another species.

A square of tape went viral in 2017

That same year, cat owners on Twitter began taping shape outlines on their floors and posting their cats sitting inside them under the hashtag #CatSquare, drawing more than 82,000 retweets by the count in Smith’s paper.

Dodman’s article took up the trend. A taped square is “a poor substitute for the real thing,” he wrote, but it may offer “some misplaced sense of security and psychosomatic comfort.”

The illusion nearly matched the real square

Smith’s study ran remotely during the pandemic, and of 561 enrollments, 30 cats finished all six daily trials. Each day their owners printed two of three shapes and taped them to the floor 2 feet apart: a square outline 8 inches on a side, a Kanizsa square of the same size, named after the Italian psychologist Gaetano Kanizsa, or the control, made of the same four Pac-Man shapes turned outward.

Owners were asked to wear dark sunglasses so they could not cue their cats, and were not told what the study was testing before or during the trials. A cat counted as choosing a shape only if it sat or stood inside with all four legs for at least three seconds.

Only 9 of the 30 ever chose a shape, and across those nine, each shape went down 36 times. The real square was picked 8 times, the Kanizsa square 7 times by four cats, and the control once.

The shape cats chose seven times and the shape one cat chose once used the same four Pac-Man pieces. The only difference was which way the cut-out corners pointed.

The authors concluded that cats “may treat the subjective Kanizsa contours as they do real contours.” In head-to-head pairings the illusion beat the control 5 to 0, a significant result, but the real square’s 3-to-1 edge over the control was not, which they called contrary to expectations. When the two squares were set side by side, cats chose the real one five times and the illusion twice, a gap the authors found not significant.

They blamed their small dataset mostly on owners dropping out of a six-day design and suggested repeating the test in controlled settings.

The 2021 study was not the first test of cats on illusory edges. In 1988 a Northwestern University team trained cats to respond to a figure defined by subjective contours and reported “the first direct evidence that a nonhuman perceives subjective contours.” Smith’s paper notes that the 1988 study used two cats.

The warmth figure rests on one disputed range

Pet-care pages that give warmth as a reason usually attach a number. Reader’s Digest puts the optimal ambient temperature for cats at “upwards of 85 degrees Fahrenheit,” Purina’s UK site says around 37 C, and a Florida veterinary practice says cats need 86 to 97 degrees F.

None of the three names a source, but all three fall inside a published range. A 2016 review of cat housing gives the thermoneutral zone of domestic cats as 30 to 38 C (86 to 100.4 F), against the roughly 22 C of most homes. In 2023 the European Food Safety Authority said that range “might have been misinterpreted and should be reconsidered.”

Neither Dutch trial was set up to test warmth. Both compared a box with no box, so they cannot say how much of the benefit came from hiding and how much from warmth. The 2019 paper, whose quarantine wards ran at 16.0 to 19.8 C and whose boxes were lined with towels, discusses hiding and never offers warmth as a reason for using the box.

Hiding is the best-tested reason a cat gets into a box. A square on the floor hides nothing, and for four cats in 2020 it did not even need its sides. Strictly FYI.