Via Schneier, we learn a bit more about cleaner fish. I'd heard of these before: little fish that eat parasites off of bigger fish, with both parties being made better off for the exchange. But it's far more interesting than I'd thought.
The cleaner fishes get to choose cooperate (clean parasites) or defect (eat the client fish's tasty skin mucous) for each client. The client fish often leaves if the cleaner chooses defect.
When working alone, male and female cleaners defect against clients about equally. But when cleaner fish work in pairs (male/female), female defection rates drop considerably. If the client leaves because the female defects, the male punishes her by chasing her aggressively, reducing her likelihood of defecting for the next client. Presumably the (smaller) female isn't as able to punish the male for defection. On the whole, average defection rates are lower for partnered cleaners, and clients seek out cleaners that work in pairs.
Here's the rather nice Nature article, October 2008.
And of course recall that Gordon Tullock invented the field of bioeconomics back in 1971.