There's a very interesting article in Nature this week by Jared Diamond about type 2 diabetes and the evolution of 'thrifty genes'. The idea is that the genes predisposing people to this form of diabetes persist in the population because they have some selective advantage. One hypothesis is that these genes have an advantage in 'feast-famine' conditions, allowing people to gain weight rapidly and lay down a lot of fat in infrequent periods of plenty so that they are then more likely to survive the subsequent famine.
This is interesting in itself, but the really intriguing thing is that there is a marked difference in diabetes rates amongst people of different ancestries: in most geographical areas, there is a large amount of variation, with rates going as high as 50% prevalence in the people of the Pacific island of Nauru. But in people of European descent (in Europe, the Americas and Australasia), the rates are very low. These low rates persist despite the fact that other predisposing factors for diabetes — a sedentary lifestyle and large amounts of excess food — were practically invented by Europeans.
Diamond proposes a possible explanation for this difference, in which famine conditions select for the thrifty genes, which then predisposes people to type 2 diabetes once food becomes widely available and in constant supply. Eventually, the genes are selected out of the population as those carrying them (without medical intervention) die of the disease. He suggests that this happened in Europe sometime after the Middle Ages, when famine became almost unknown. I'm not sure that I agree with all the arguments, but it's an very interesting idea, and Diamond proposes some tests which might provide more support — always a good sign in a theoretical article.