Tool-toting New Caledonian crows may be smart, but are they schemers?

The cunning birds can sculpt twigs into hooks and wield multiple tools in succession to obtain an otherwise unreachable snack. This kind of behaviour has prompted some scientists to conclude that the birds plan their actions ahead of time to achieve a goal.

However, new experiments led by Joanna Wimpenny and Alex Kacelnik at the University of Oxford cast doubt on that conclusion.


Multi-tool users?

The researchers tasked seven crows with snagging a snack - in this case a morsel of pig heart - trapped down a clear tube. To get this treat, birds needed to use a short stick to ferret an intermediate stick out of a different tube, which they used to obtain a third stick, which was long enough to reach the food.

Positive reinforcement - where a good result encourages an action - may explain why any animal uses a tool to obtain or process food: chimpanzees that fish termites out of the ground with sticks or crack nuts with rocks, for example. But explaining how an animal learns to apply one tool to another tool is more difficult, Kacelnik says.

"If you think of picking up a tool, addressing it to another object - not food - and using the object to find food then you have an even more improbable chain of events," he says. If learning by reinforcement was the explanation, then the chances of using three objects in succession are "ludicrously small".

Old hands

Despite these long odds, four New Caledonian crows passed Kacelnik's test. However, these birds had previously practised one part of the sequence: retrieving stick tools with their beaks and using the tool to get a treat, and even then they varied in the amount of attempts they took to complete the task.

On the other hand, two birds that had never rehearsed the basic task struggled to complete even one step. While a third untrained bird learned to extract food in a simpler task, after 21 failures.

These errors suggest that the crows understand some aspects of the test, but probably not the full complement of sequences required for success, Kacelnik says.

Similarly, crows tended to swap smaller sticks for larger ones - but not always. If the birds had formulated step-by-step plans, they would not have made such errors, Kacelnik says. "The fact that an animal is statistically better than random should not be confused with the animal being perfect."

Journal reference: PLoS-ONE