I think it's certainly worth mentioning that there's a pretty huge difference between the types of discrepancies one sees between different manuscript sources at the beginning and at the end of the eighteenth century. And that even at its most extreme there's usually not that much difference between sources unless they're outright recompositions.

Great picture, Ollie!  And very apt...

Corelli is a bit earlier than Vivaldi, but provides an interesting example of ornamentation.  His Opus 5 violin sonatas were published in 1700.  Ten years later, Estienne Roger published a version of Opus 5 that included embellishments supposedly supplied by Corelli himself to demonstrate how he played the slow movements.  Only the slow movements had any ornaments or decoration - and very elaborate they are indeed.  There is also a slightly later ornamented version of the same set by Geminiani - again, only the slow movements are embellished.  That's not to say there wouldn't have been any ornaments in the fast movements - cadential trills, turns, appogiaturas and similar stock ornaments would (probably) have been second nature (and these sorts of ornaments were well described in treatises of the time).  But outrageously elaborate wanderings were confined to the slow movements.

I don't know of any other pieces that received similar published treatment.  And here's a little gem I just found on Grove:  "William Hayes, Professor of Music at Oxford, was contemptuous of Geminiani’s habit of giving specific ornament and phrasing instructions to performers, claiming that he was ‘paying his Brethren of the String but an ill Compliment’.  The attitude that, at best, explicit markings are patronizing may help to explain why they are found in relatively few scores."

On the other hand, manuals on how to do ornaments and divisions had been around for a couple of centuries.  Sylvestro Ganassi's Opera intitulata Fontegara was published in Venice in 1535, and gives page after page instructing the recorder player how to decorate cadences as well as rising and descending seconds, thirds, fourths and fifths.  Similar manuals existed for cornett players.  Much 17th century music looks very simple on the page, and lends itself easily to this style of elaborate embellishment.

Biber is a significant exception to this.  His music as written is already quite elaborate - and usually when it's not, it seems to be part of the pacing in building up towards more elaborately written things later.  One explanation is that he was a phenomenal violinist, and wrote what he was able to do.  Ornamenting things that are already ornamented is a bit superfluous, so there's less room to play around with the notes as written.  On the other hand, the written-out decorations are loads of fun to make sound spontaneous.

I think this may be the case with Bach as well.  He was a champion improvisor himself, and was master of both keyboard and violin, and often there isn't room for much more from the performer in his work.  But a lot of the bends and twists in his work inhabit a magical space between harmonic structure and decorative elaboration, and (I think) are best approached as if spontaneous, even though very few people indeed would be able to take things in the directions Bach does...

I have a dim recollection of hearing about some sort of engraved device for automatically playing an organ (similar to a piano roll) that was done of a Handel sonata in the mid 18th century.  Someone attempted to transcribe and reproduce the ornaments in (again) a slow movement.  It was really, really, really weird and elaborate.

...Or not, perhaps. After all there are two examples of cadenza-like things being written down in the Brandenburg set: the recorder ornamentation near the end of the slow movement of no. 4 and the enormous harpsichord cadenza in the first movement of no. 5. And 'Adagio' at that time didn't necessarily have to refer to a separate movement but could just be an indication corresponding more or less to 'rit', as at one point in the third movement of no. 1. The oboe figuration in the slow movement of no. 1 is also the kind of thing that could just as well have been improvised. So perhaps just the two chords separating the movements is indeed what he meant? (Certainly I find that entirely persuasive from MAK and Café Zimmermann, to take but two.)

Oh, I agree with you...   why write two chords unless that's what you intend?  (I can see the reasoning that there might have been more chords of which these were the last two, however).   But over those chords there has to be "something"...  from an attractively elaborate chord-realisation through to a cadenza?   To me it would seem odd that in a Concerto featuring stringed instruments, that one of them would not have played that cadenza?  (Why bring the continuo instrument to the fore merely for that moment?).  The "presentation" quality of the MSS mitigates any kind of Rossinian "no time to write it down!" explanation, IMO?