Raccoons are smart, but not as smart as crows

October 5, 2017 • 9:45 am

. . . at least when in comes to water displacement experiments.

Via ZME Science and reader Ant, I learned of a new study trying to see if raccoons (Procyon lotor) could solve the Archimedes puzzle. This is the puzzle, formalized by Aesop in his fable “The crow and the pitcher“, that determines whether an animal can figure out how to use water displacement to fetch a treat. That is, if you have a floating treat that’s in a water-filled vessel, but the water level is too low to fetch the treat, can the animal figure out that adding stones to the water will raise the level, making the prize accessible?

It’s long been known that crows can learn this with great skill, but a group of researchers in the U.S. wanted to see if raccoons could do it, too. Their hypothesis was that yes, the beasts could. They did a study on eight raccoons (four wild-caught and later released, four reared in captivity) to see if these famously clever carnivores could also figure out water displacement. The authors’ results, published in a paper in Animal Cognition (reference below, access is free), are a mixed bag: a few raccoons could figure it out, but #NotAllRaccoons. And it was a mess, because these animals didn’t cooperate well, playing with the stones, messing about, and even overturning the heavy apparatus to get the treat.

I’ll be brief as the results, while interesting, aren’t particularly stupendous. The authors had a two-part design, with each part subdivided into sub-parts presented in succession.

A. Raccoons were given stones and a marshmallow floating in a deep, half-meter cylinder partly filled with water. If they didn’t succeed in learning to use the stones, they went on to part B.

B. Raccoons were helped out by balancing stones on the lip of the tube with food placed on top the stones. Their messing about and getting the food could cause the stones to fall into the water, perhaps helping them learn what to do.

C. Raccoons completing part B were then given stones lying about on the cage, with the aim of seeing if they’d learned how to use them after being exposed to part B.

All raccoons who learned to drop stones into the water then progressed to “Phase II”, which had four parts. As far as I can see, all raccoons in Phase II were subject to all four sub-studies:

D.  Raccoons were given three big stones and three little ones. Could they preferentially use the big stones?

E. Raccoons were given two tubes with treats: one with water and one with corncob litter. Were they smart enough to realize that the stones would work only with the water?

F. Raccoons were given six tennis balls instead of stones; three of the balls were heavy and would sink, displacing water, while the other three would float and were useless. Would the raccoons be smart enough to use the heavy balls?

G. Raccoons were given a steel cup with a handle that they could use to scoop out the marshmallow bits. Could they learn to use it?


One of the eight raccoons wasn’t interested in the task, and was removed from the trials.

No raccoon succeeded in part A: figuring out on its own how to use the stones.

In part B, four raccoons accidentally knocked the stones into the tube and retrieved a treat.

In part C, just two of the five animals subject to part B learned to drop the stones into the tube, and thus progressed to Phase II. Another raccoon messed around and got the treat this way:

“During final trials, Raccoon 22 innovated a unique solution by gripping the inner rim of the apparatus with her forepaws and, while rocking her body back and forth, overturned the entire apparatus and retrieved the reward.”

So the sample size for Phase II was only two raccoons—not enough to say much.

Neither raccoon learned to use the big rather than small stones in part D.

In part E, only one raccoon preferentially dropped stones into the water; the other dumb one kept dropping stones into the corncob litter.

In part F, neither raccoon preferred the dense balls to the floating balls, and so didn’t learn to get their treats that way. But both were observed to push the floating balls down into the water, splashing up bits of marshmallow that they could retrieve.

Finally, neither raccoon learn to use the cup to scoop out marshmallow bits, though each, on just one occasion, dropped the cup into the water and fished out marshmallow bits with it before the cup sank.

The upshot: Raccoons are either dumber than crows or weren’t engaged in the task. Only two out of the eight learned to properly retrieve the marshmallows.

My alternative hypothesis (which is mine): Raccoons don’t like marshmallows all that much, and weren’t willing to go to much trouble to get them. As the researchers mention below, some raccoons “did not seem to be goal-oriented.”

The rather long discussion goes into reasons why, despite the authors’ predictions, the raccoons were recalcitrant. For some reason I found this part of the discussion hilarious (my emphasis):

The exploratory, tactile nature of raccoons may have confounded their performance in the Aesop’s Fable paradigm. For example, during Phase II the behavior of Raccoons 29 and 40 did not seem to be goal-oriented, in the sense described in many other Aesop’s Fable studies (e.g., Bird and Emery 2009 ). That is, they did not drop the exact number of stones necessary to retrieve the reward and continued dropping stones and exploring experimental materials after the reward had been retrieved. We recorded many instances where the raccoons washed the stones/objects in their water dish, buried the stones/objects in their litter box, carried the stones/objects into their den box, and seemingly played with the stones/objects for long periods of time.

In other words, perhaps they were more interested in other things than getting marshmallows, or maybe they’re less food oriented than are crows. Who knows? At any rate, I end by showing two videos taken from the paper along with the descriptions given:

Video footage from Raccoon 29’s eleventh tool use trial. He moves the cup around the opening of the tube with his paws and mouth for several seconds before releasing it into the tube. He then quickly grabs the handle of the cup before it sinks, and retrieves a piece of marshmallow as he pulls the cup out of the tube.

Video footage from Raccoon 29’s second substrate trial. He first drops a stone into the water tube, retrieves and eats a piece of marshmallow, then selects a second stone and drops it into the corncob tube. After he is unable to obtain the reward from the corncob tube, he returns to the pile of stones, makes a selection, and heads toward the water tube.

SUCCESS! (Note how skinny the tubes are!). But then this one goes over and drops tubes into the corn litter, wasting its time.

h/t: Ant


Stanton, L., E. Davis, S. Johnson, A. Gilbert, and S. Benson-Amram. 2017. Adaptation of the Aesop’s Fable paradigm for use with raccoons (Procyon lotor): considerations for future application in non-avian and non-primate species. Animal Cognition, online.  https://doi.org/10.1007/s10071-017-1129-z

24 thoughts on “Raccoons are smart, but not as smart as crows

  1. I think an animal that just turns over the receptacle rather than laboriously filling it with stones is pretty smart. Sort of like Indiana Jones just shooting the swordsman.

    1. Agreed, raccoons don’t avoid water, they enjoy it. If they were really hungry they’d have found a wat to rip those tubes down judging by what Ive seen them do to trash cans.

  2. This is great:

    One of the eight raccoons wasn’t interested in the task, and was removed from the trials.

    Raccoon: “Umm, why are we doing this moronic exercise? Could you PLEASE let me back out into the woods where you found me? Thanks a lot. No hard feelings.”

  3. I wonder if a problem might have been raccoons’ fabled fondness for collecting baubles. When I was a young teen I did a “report” on raccoons for school. One interesting thing that has stuck with me for some reason was a type of trap I read about while researching for that report.

    The trap took advantage of raccoons’ supposed obsession with collecting interesting baubles. The trap was a section of log with a hole drilled in one end. Large nails were driven into the log, placed and angled so that a raccoon could reach into the hole with an open paw / hand but would not be able to pull their paw / hand out if their hand was closed around an object. Suggested “bait” for the trap was a shiny object like a large washer, a polished stone, etc. Supposedly raccoons are so obsessive about such baubles that this trap would work because the raccoon would be unwilling to give up the bauble in order to get its hand back out of the hole.

    Now, I’ve really no good idea about how accurate the claims about that trap or raccoons’ obsessive bauble collecting behavior is. But if there is some truth to it that may be why the raccoons were slow to figure out how to, or decide to, use objects like stones and tennis balls to drop down a tube to get the marshmallows. They may have been more interested in the cool objects.

    1. Crows have the same obsession. They like “Shiny Things” according to the Tom Waits’ song of the same name.

  4. These tests on (non-human) animal abilities can be difficult to compare across species. While in college I met a fellow student who worked in a lab that ran tests like these. He described one experiment where they tied squirrels and d*gs to long leashes that were wrapped around obstacles. The task was for the animals to unwind the leash so as to be able to reach some food placed just out of reach. The squirrels easily mastered the task of re-routing the leash around the obstacles but the d*gs, predictably, were abject failures. But no one would claim that squirrels are more intelligent than d*gs. Sometimes these tests are testing something other than intelligence. That may well be the case here.

      1. I bet you do. 😉

        My recollection is that they thought the squirrels could do the task better than d*gs because squirrels spend their lives navigating complex pathways through trees to get where they want to go – they are “wired” for that kind of task. When they gave them a different test of intelligence (unfortunately I cannot now remember it), the dogs easily mastered it but the squirrels were stumped. The lesson I learned from that was that it can be difficult to compare performances if the tests are not very carefully vetted.

  5. I remember reading once about an experiment where a chimpanzee was given a long stick to see if it would use it to knock a treat loose from a high shelf. The chimp outsmarted the testers by standing the stick on end and climbing it to reach the treat, a solution they hadn’t considered.

    1. I saw a TV show on Honey Badgers who were able to defeat every single attempt to keep them in their open air, high walled enclosure, including piling up stones high enough to reach the top, or pulling limbs off trees to build a ladder or once they combined the two – putting a limb on top of a pile of stones they collected. They were also able to figure out how to unlock the doors. Only a permanent overhead fence and padlocks were able to defeat them.

  6. For some reason I read the headline as “not as smart as cows.” And I thought wow, that’s a pretty low bar for the raccoons to fail to clear.

  7. Only two out of the eight learned to properly retrieve the marshmallows.

    Or maybe “in only two out of eight cases did the researchers properly predict the raccoons’ behavior.”

    Personally, I think Raccoon 22 should be given command of the Enterprise.

  8. The raccoons who successively completed the tests should be bred. Their offspring should be subject to a battery of new tests. Raccoons who succeed at those new tests will then be bred. The process will continue until we have a species of super intelligent raccoons, who will then run for congress.

Leave a Comment

Your email address will not be published. Required fields are marked *