cnd, thank you very very very much for that! I'll look it over and report back if I think I have anything interesting to add.

You might remember a thread I started where I made a case for seeing "Shenandoah" as always metrical, and quite necessarily so, yet in a different meter than most people imagine it. Part of that "theory" is that the unexpected meter confused collectors, so many concluded it wasn't metrical.

Anyway, I was vague in my last message, partly out of laziness and partly because I'm working on writing a piece right now and I've sort of bored myself talking about it, ha. I apologize.

But to clarify: I was referring to a very specific difference in melody versions, in pitch (ignoring the meter issue). It's the difference in the two notes on the words "hear you" (as in "Shenandoah, I long to hear you"). According to my research, all documents of the tune (published, manuscript, field recordings) before 1920 are consistent in having a certain two notes at that point. Then, in 1920 (but more prominently in a 1921 publication), the chanty collector/editor comes out with a melody where those two notes are different. (I'm calling the former Version A and the latter version B.) Since Terry's published score was a rare case of an accessible, full text with full piano (and violin!) accompaniment, suitable for concert performances, concert singers in the 1920s and beyond took it as their guide. With time, this Version B became the presumed standard way to sing "Shenandoah" (even while continued fieldwork turned up sailors who remembered the ~original~ Version A).

So what I was getting at is that the Lomax collection, 1934, has the Version B notes. Either this was some mistep by Lomax, under the influence of thinking Version B was more correct (i.e. in the wake of the Terry phenomenon) OR it reflects the fact that Version B actual was around (as remembered by Spalding) before the Terry phenomenon. The latter, was certainly possible, is something I find rather incredible. There's just so much evidence (relatively speaking!) that Version A was the thing. So while it's out of order to accuse either Lomax or Spalding of playing hanky panky with the melody, to me that checks out as more likely.

I'm biased, but that being the case it inclines me to look even more suspiciously upon the "Tommy Tompkins" thing. In other words, I figure that if Spalding's text (or Lomax's mediation of it) was less than totally original, then the melody could be, too.

As to meter -- and this is an aside to my other discussion about meter -- seeing it in 4/4, as it appears in Lomax is rare (and thus also divergent from the bulk of evidence). Lomax simply putting it in 4/4 as a result of struggling with the meter would be par for the course with other collectors and thus pretty unremarkable. Yet if Spalding definitely gave a tune in 4/4, because that goes against what I believe to be the fact that the sailors' chanty was not in 4/4, it would be another way the hypothetical "cavalry version" is distinctive.

In sum, the preponderance of evidence, at least how I interpret it, is saying that the historical sailors' chanty was neither in 4/4 meter nor did it have the "Version B" melody figure. This makes the Lomax/Spalding example really problematic. Either these are errors/contrivances that make Lomax/Spalding unreliable, or, if we presume it to be reliable, it's an irksome exception that muddles the whole picture.