This is a very interesting discussion!
I think there are two separate questions.

(1) the distinction between a binding fragment or endleaf, and the composite manuscript consisting of several originally separate manuscripts bound together. Personally, I have been happy to use  `<msPart>` for all these cases;  one can use `@type` to provide a more precise indication of what the part is. For example, type="pastedown", type="endleaf", type="booklet" and so on. It might even be possible to agree on a controlled vocabulary...

(2) The question of describing a manuscript which is known to have leaves elsewhere. I would argue against Pietro that one might actually want two descriptions here - one of the surviving MS. and one of the reconstructed MS. Would it also be possible to describe BNF et 45 using only `<msFrag>`?