Last week I published a short article in Molecular Ecology on evidence for natural selection. It has proven difficult to show natural selection occurring in real time in wild populations. New approaches may help, and these are being pioneered in studies of Soay sheep. While commenting on these new approaches, I make several general points about the evidential case for natural selection.
Perhaps the more broadly interesting of these is a critique of the argument by analogy to natural selection. I suggest that, although widely used, the analogy has severe limitations. You can read this critique from the second to the seventh paragraph of my article, which is available open access here.