Are crows mind readers? Recent studies have suggested that the birds hide food because they think others will steal it -- a complex intuition that has been seen in only a select few creatures. Some critics have suggested that the birds might simply be stressed out, but new research reveals that crows may be gifted after all.
Cracks first began forming in the crow mind-reading hypothesis last year. One member of a research team from the University of Groningen in the Netherlands spent 7 months in bird cognition expert Nicola Clayton's University of Cambridge lab in the United Kingdom studying Western scrub jays, a member of the crow family that is often used for these studies. The Groningen team then developed a computer model in which "virtual jays" cached food under various conditions. In PLOS ONE, they argued that the model showed the jays' might be moving their food—or recaching it—not because they were reading the minds of their competitors, but simply because of the stress of having another bird present (especially a more dominant one) and of losing food to thieves. The result contradicted previous work by Clayton's group suggesting that crows might have a humanlike awareness of other creatures' mental states—a cognitive ability known as theory of mind that has been claimed in dogs, chimps, and even rats.
In the new study, Clayton and her Cambridge graduate student James Thom decided to test the stress hypothesis. First, they replicated earlier work on scrub jays by letting the birds hide peanuts in trays of ground corn cobs—either unobserved or with another bird watching—and later giving them a chance to rebury them. As in previous studies, the jays recached a much higher proportion of the peanuts if another bird could see them: nearly twice as much as in private, the team reports online today in PLOS ONE.
Then came the stress test. First, Thom and Clayton gave the jays trays with the ground cobs but no food to hide in them—a so-called "sham" session. Then, in a second session, they gave the birds new hiding trays and bowls of peanuts to hide. When the jays were done, the experimenters removed the trays and stole all of the peanuts. Finally, after a short break, the researchers gave each bird yet another round of food, a new tray to hide it in, and one of the trays it had seen earlier: either the sham tray or the ransacked "pilfer" tray. The jays had 10 minutes for recaching.
If the Groningen model was correct, Thom and Clayton argue, the stress of discovering that food was missing from the pilfer tray ought to drive jays to cache more peanuts than those presented with the sham tray. In fact, there was no difference, even though corvids have excellent memories for hidden food and remarkable abilities to find it again. The hypothesis that jays have theory of mind remains on the table, Thom says.
Thom and Clayton have "definitely shown that scrub recaching is not as simple as the [Groningen] model presents it," says Elske van der Vaart, lead author of the Groningen team's earlier report, who is now at the University of Amsterdam. But she argues that there is still room for doubt about what the results mean. For example, the sham condition—in which the jays had no food to cache—could have stressed the birds as much as the stolen peanuts in the pilfer condition did.
Amanda Seed, an animal cognition researcher at the University of St. Andrews in the United Kingdom, says the Groningen model's failure to predict the birds' caching behavior in the new experiments could "bring the model down like a pack of cards." But researchers still have to rule out other possible explanations, she says. For example, the birds given pilfered tray may have noticed the missing peanuts too late to affect their overall caching rate, or they may have spent much of their time looking for the missing nuts instead of hiding the new ones. The Cambridge and Groningen groups are planning more work with both real and "virtual" birds to see what is really going on. "I applaud them for rising to the challenge," Seed says.