"Generalized reciprocity" present in rats

Rats show a type of "generalized" altruism:

Rats that benefit from the charity of others are more likely to help strangers get a free meal, researchers have found.

This phenomenon, known as 'generalized reciprocity', has only ever been seen before in humans. A good example, says Michael Taborsky of the University of Bern, Switzerland, is what happens when someone finds money in a phone box. In controlled experiments such people have been shown to be much more likely to help out a stranger in need following their good luck.

In humans, such benevolence can be explained by cultural factors as well as by underlying biology, says Taborsky. But if similar behaviour can be found in other animals, he reasons, an evolutionary explanation would be far more likely.

To test for this behaviour in animals, Taborsky trained rats to pull a lever that produced food for its partner, but not for itself. Rats who had received a free meal in this way were found to be 20% more likely to help out an unknown partner than rats who had received no such charity.

Taborsky believes this behaviour isn't confined to just rats and humans. "I'm convinced generalized reciprocity will be very widespread and found in many different animal species, as our study suggests that an underlying evolutionary mechanism is responsible."

"Generalized reciprocity is certainly underappreciated in animals," notes Laurent Keller, an evolutionary biologist at the University of Lausanne, Switzerland. But with so little data in this area, he adds, "it is very difficult to make predictions about how prevalent it's likely to be."

This generalized reciprocity is distinct from the reciprocal altruism you have probably heard of before because in this case the animal that the rat helps is not one that it is necessarily familiar with. The rat senses that it is in a millieu where helping is prevalent, and it contributes to that millieu regardless of its previous relationship with the recipient.

Taborsky distinguishes this from reciprocal altruism with a known animal by showing that rats are 50% more likely to help an animal that had helped them in the past (it's at the end of the article).

This stuff is interesting for two reasons. One, I had been under the impression that the cognitive machinery necessary for reciprocal altruism was not present in rats. Apparently I was wrong. Two, this generalized reciprocity may be present even in more simple creatures than rats because it does not require recognition of the other animal in order to work. It might represent the most evolutionarily old form of reciprocity.

How would such reciprocity evolve? Last year I posted an article about how cooperation could emerge without cognition provided the animals who chose to cooperate were isolated. Say a set of animals who shows more tendency to generalized reciprocity becomes geographically isolated. This research would suggest that reciprocity is not only evolutionarily stable would be highly selected for.