A bit of future-think lets jays cooperate. (Trust That Bird?).

A blue jay will cooperate with a buddy for mutual gain in food despite opportunities to betray the partnership, according to a new laboratory study.

Such cooperation among animals had remained elusive and controversial during decades of scientific studies, explains David W. Stephens of the University of Minnesota, TWin Cities. Now, he says, he has demonstrated why. The earlier work hadn't taken into consideration the timing of the benefit for cooperation, he and his colleagues contend in the Dec. 13 Science. The partner's reciprocity also influences a bird's choice to cooperate, they find.

This work will push behavioral biologists to think about the timing of rewards in other contexts, such as foraging, aggression, and mating, predicts Eldridge S. Adams of the University of Connecticut in Storrs. "I think it's quite clever," he says.

Many examples of animal cooperation in the wild turn out on close examination to be cases in which no partner benefits unless all act together, Stephens says. However, he's more interested in the situations where there's an incentive for cheating. Scientists have modeled these with a scenario, called the Prisoner's Dilemma, in which human conspirators under interrogation can either protect their partnership or rat on their ally. The model predicts that an animal in such a setup will cooperate when it repeatedly encounters a partner that reciprocates cooperation.

"The animals in the lab always seem to cheat," says Stephens. "There's this compelling question of what's gone wrong with this elegant model."

He says that he's long suspected that experimental animals don't cooperate because of temporal discounting, a tendency to devalue the future possibilities in favor of immediate gain. In a variety of other tests, "animals often behave as if they care only about the next few seconds," he says.

He and his colleagues have now designed a way to coax animals to focus on long-term consequences. The researchers put a test bird in its own enclosure, with a single perch at the back and two in the front. Its partner had an identical setup next door. The researchers trained the birds to fly from the back perch to a front one, and opaque partitions kept the birds from seeing each other until they reached a front perch.

Landing on one perch earmarked a small food reward for a bird. Landing on the other choice, the cooperation perch, assigned a larger food reward for the neighbor but none for the chooser. The innovative twist of the setup was a clear box that accumulated each bird's winnings and released them either immediately or after four rounds of the game.

One bird, the test subject, could choose either perch, but the researchers controlled the choice of the other bird during 1,000 rounds of the game.

When the researchers let animals have their rewards immediately, the test bird reduced its cooperation no matter where the neighbor landed. When the researchers kept the neighbor steadily uncooperative, the test birds likewise reduced cooperation. However, the researchers did see sustained reciprocity when they delayed rewards and kept the neighbor cooperating, too.