Useful information for training your cephalopods


Everyone knows the story of Konrad Lorenz and his goslings, right? It was a demonstration of imprinting: when young animals are exposed to a stimulus at a critical time, they can fix on it; Lorenz studied this phenomenon in geese, which if they saw him shortly after hatching, would treat him like their mother, following him around on his walks. Similarly, many animals seem to experience sexual imprinting, where they acquire the sexual preferences that will be expressed later on.

I just ran across a charming short letter about imprinting in cephalopods, and somehow the story seems so appropriate. Imprint a young, freshly hatched cuttlefish on something, and they don’t treat it like Mom, and they don’t later want to mate with it—they want to eat it. Lorenz was lucky he was working on birds rather than cephalopods.

The experiment is straightforward. Cuttlefish normally prefer to eat shrimp over crab. If, the day after hatching, small crab are put in the tank with the hatchlings for at least two hours, and then removed (the crabs are not eaten), then 3 days later when tested again, the cuttlefish will prefer to kill and eat crabs over shrimp. The procedure is very specific: they have to be exposed to crab for at least two hours, within 2 hours after sunrise on their first day after hatching.

The paper has a good, succinct description of why many animals would have this mechanism:

Precocial animals, like domestic
chicks and cuttlefish, which are
independent within hours of hatch
or birth and which receive no
posthatch parental care have
two options for acquisition of
information: bring it into the world
with you (unlearned preferences
for food, sexual partners and so on)
or pick up the information as you go
(trial and error learning). Imprinting
allows something in between:
a certain degree of flexibility in
response, useful for learning
information for which the timing is
likely to be predictable—food
seen in first few hours of life,
sibling/parents seen during
juvenile stages—but in which
specifying the exact details of
the experience is not useful.

An evil man could think of many nefarious things to do with this bit of information, I think.

Healy SD (2006)Imprinting: seeing food and eating it. Curr Biol.16(13):R501-502.


  1. says

    The hard part will be getting our leaders to take a swim in the giant squid maternity ward for at least two hours, within two hours after sunrise on the first day after hatching.

    Note to Homeland Security: that was a joke. I do not and would not ever advocate feeding our leaders to giant squid.

  2. Alex says

    If it works for sharks, we could do that and “put frikken lazer beams on their heads!!….rrrriiiiigght”.

  3. Goldensteinian says

    nd ths ll prvs tht thr s n Gd nd tht PZ Myrs hs ll th nswrs.

    Hlp m pls, dn’t knw wht t d nw.

    Knss Ctzns fr Scnc hs bnnd m nd nly tw ppl shw p t Brns nd Nbl fr my dscssn grps.

    Wrs, my spply f Chck Trcts s rnnng lw.

    fnd mldw n th grg bt hv 14 cps f Th Pttn Mnd fr sl nd lrg cllctn f Slvr g Cmcs.

  4. NJ says

    Hey, PZ could stay ahead! All he needs to do is squirt out an inky jet of… of…

    OK, not going there.

  5. llewelly says

    The next step is to combine the two experiments. Expose newly hatched chicks and cephalopods to each other, so the chicks will imprint on the cephalopods and the cephalopods will imprint on the chicks. Then observe the results …

  6. Steviepinhead says

    Well, I can’t get Warren’s link to work, so I don’t know if he’s well ahead of me here, but this study could certainly explain all those rather kinky Jampanese prints…

  7. Steviepinhead says

    Uh, make that “Japanese.”

    And, yes, there was a distant echo of a pun there: imprints ==> prints. But whatever virtues might otherwise have been, plowed under by typo now are.

  8. says

    PZ: been to the Seattle Aquarium lately? They have a guy and a girl octpus in a particular display, separated by some sort of metal wall, bolted in place. The guy figured out how to unscrew the bolts. Also, he apparently took a particular dislike to a guard who’d shine his flashlite on him at night. Once the guard walked by during the day and — so they say — the octopus squirted ink at him.

  9. Sophist says

    PZ could still lead a trail of octopi a la Lorenz, but only as long as he could stay ahead of them.

    I’m fairly certain that PZ could outrun an octopus. I mean, they’re not really known for their fleetness of foot, are they?

    And as long as we’re talking octopi

  10. don says

    In Gravity’s Rainbow it was Pavlovian conditioning (I think) rather than imprinting, but Octopus Grigori was definitely trained and used for nefarious purposes.

  11. strewelpeter says

    An evil man could think of many nefarious things to do with this bit of information, I think.

    Isn’t that what the advertising industry does?

  12. Carlie says

    The next step is to see how sensitive the visual cues have to be. Is it simply to the level of gross features, like red for the stickleback fish, or can they distinguish body shapes, could they distinguish between a blow-up doll and an actual person, or could they tell the difference between Bush and Kerry?

  13. ajay says

    Isn’t that what the advertising industry does?

    With squid?? No. Well, I don’t think so. Come to think of it, how would I know? I doubt we read the same magazines as each other.

  14. says


    Well, I can’t get Warren’s link to work, so I don’t know if he’s well ahead of me here, but this study could certainly explain all those rather kinky Jampanese prints…

    Nahh, it was a picture of Ratzi.

  15. says

    Somebody beat me to the Japanese reference, but my first thought (dirty, dirty thought) was in regards tentacle Hentai.

    Are these based on Japanese women swimming before breakfast?