From The Economist:
Until now, experiments to try to determine the biological basis of beauty have been of the please-look-at-these-photographs-and-answer-some-questions variety. Some useful and not necessarily obvious results have emerged, such as that one determinant of beauty is facial symmetry.
But what would really help is a breeding experiment which allowed the shapes of faces to be followed across the generations to see how those shapes relate to variations in things that might be desirable in a mate. These might include fertility, fecundity, social status, present health, and likely resistance to future infection and infestation. Correlations between many of these phenomena and attributes of the body-beautiful have, indeed, been established. But in a pair-forming, highly social species such as Homo sapiens, you also have to live with your co-child-raiser or, at least, collaborate with him or her. So other things may be important in a mate, too, such as an even temper and a friendly outlook.
It would be impossible to do such a breeding experiment on people, of course. But as Irene Elia, a biological anthropologist at Cambridge, realised, it has in fact been done, for the past five decades, on a different species of animal. Dr Elia has published her analysis of this experiment in the Quarterly Review of Biology. The animals in question are foxes.
More here.