Steve Gardham wrote: If it is the case that a substantial number of versions of the shorter ballad have significant crossover with the old ballad, then a detailed study is needed. Is it possible that some versions of HM have been deliberately crossed with SAB within the last 2 and a half centuries? There are plenty of examples of mixing of ballads by, shall we say, literate people, and then those mixtures have gone into oral tradition.

It's perhaps worth noting that Steve Roud agrees with me; he lists both AB and HM as #104, and lists the versions with BOTH Child numbers.

Which is not to disagree with you. The two versions did not diverge enough to become separate songs (IMHO), but they did diverge enough that different versions could cross-fertilize.

Sorting it out, though, would take somebody doing a major collation project. Bronson, e.g., has 62 versions of the two collectively, and there are 288 items in the Roud Index. Some of those are non-meaningful (e.g. the two citations of the Ballad Index), and some are duplicates, but I did my best to count the different informants cited by Steve R. (not easy, because different books may call an informant by a different name), and got 111 different informants plus the anonymous ones. Some of the informants (e.g. Sam Larner) actually knew multiple versions. So I suspect we have 125-150 different texts to deal with. That's going to need somebody willing to do a substantial paper to figure it out. I have the background (I'm the one folk song scholar, these days, who understands stemmatics :-), but I don't have the time!