Mudcat doesn't offer a "like" facility, but it it did I would be liking most of the recent posts.

One big difference between biological evolution as concisely described by Pamela and that of folk songs is that the population of song S in generation N in a particular locality was often only a single specimen or, at most, a few. So there would have been very limited scope for selection of one version rather than another. Rather, selection certainly did happen between one song and another, some becoming widespread, some being collected only a very few times, and many broadsides falling by the wayside long before a collector arrived, if indeed anyone ever sang them. (Selection between versions is more relevant nowadays when those with sufficient interest can have access to many versions including those from the various collections and recordings by revival performers. But a major determinant now of which version gets learnt and performed by new singers is not any of its inherent traits but who recorded it, e.g. Joan Baez, Martin Carthy, Nic Jones ...).

As in biology, the source of variation is imperfect heritability: the singer (or in the case of the words the broadside printer) makes changes intentionally or unintentionally.