Under the signature Musica manuscripta 3224, the Bayerische Staatsbibliothek has a group of eight fifteenth-century musical fragments known to their friends as ‘MuL’. They came to public notice in stages: Julius Joseph Maier included four of them in his 1879 catalogue of the library's music; Johannes Wolf wrote about two more in 1904; and the last two had to wait until Helmut Hell revealed them in 1983.

But even the first four leaves were documents of major interest, not least because they included the folio numbers ‘29’ and ‘102’ written boldly in the middle of the upper margin in a script that plainly matches the uniform ruling, music and text hand throughout those leaves. So it was a substantial manuscript written in full-black notation on parchment.

The leaves known by 1904 added new excitement because they contain one of the most tantalising compositions of the fifteenth century, Du Fay's Iuvenis qui puellam, his comic parody of a legal disquisition, which sadly breaks off half way through but has the ascription ‘Decretalis Guillermus dufay’. What we cannot know here is whether the patent absurdity of the text and its music extended to the ascription, that is, whether we should take it as proof positive that Du Fay had a legal degree at the time of copying. If we can take it solemnly, it means that the manuscript was copied after January 1436, at which point we know that he still had no degree. That is a bit later than one would have guessed from the full-black notation and the repertory, but we can let that go.

As a result of the 1983 additions, we have a continuous run of folios from f. 102 to f. 106 – not all complete but quite enough to show exactly what was on those pages, namely Latin-texted non-mass pieces, two of which seem to be Latin contrafacts of songs that may have begun their life with English text. But the new f. 104 added an important detail, missing from the rest, namely a pencil annotation to the effect that this leaf, at least, came from a book that had belonged to the Bavarian monastery of Weihenstephan.

Now the recent research project of Alexander Rausch and Robert Klugseder in Vienna has revealed four more leaves – or rather, two bifolia, mainly occupied by Magnificat settings but including a fascinating new Veneto-style motet and bits of Du Fay's Flos florum with readings that look better than the other three sources that we have for that piece. Margaret Bent immediately recognised that these new leaves belong to the same set. With that information, Klugseder then followed up the clue left hanging by Hell's article of 1983, namely to search among the Weihenstephan books in Munich for other traces of MuL. This had some fairly spectacular results, not in the book mentioned in the annotation (which he shows had contained ff. 103–6 of MuL) but in the volume from which Maier's original four leaves had been taken.

First, in the latter binding he found a further sliver from the top of f. 22v, containing an ascription to ‘Raynaldus de lantins’ for the Credo of the mass cycle ascribed in two other sources to Arnold de Lantins. Elsewhere among the fragments of MuL there is a fauxbourdon setting of Ut queant laxis credited to ‘Ray. de lan’. Until now Ray has been considered yet a third composer called ‘de Lantins’, a composer known from just this one small piece but all the same getting an honourable mention in both Grove and MGG. Now Margaret Bent plausibly suggests that ‘Raynaldus’ and ‘Arnaldus’ are one and the same. We have another piece by Arnold, and an independent ‘Ray’ disappears in a puff of smoke.

Second, the same sliver has on its recto side the clearly written folio number ‘22’ precisely matching the other centrally written foliations on all the leaves for which we have the top. That makes reconstruction of this portion of the manuscript fairly clear:

This pattern is familiar from other sources of the time, particularly the two in Bologna: mass movements in a not particularly logical sequence, with songs added in any unused space.

Sadly, it is not possible to move on from there to reconstructing the gathering. Klugseder has demonstrated that all the bifolia were separated before being reused for bindings. And the assembly of the original manuscript did not follow ‘Gregory's Rule’, which states that flesh-side should face flesh-side and skin-side should face skin-side: by that system, an odd-numbered and an even-numbered sheet as close as ff. 22 and 29 should not both have skin-side rectos. So there is no real hope there.

Third, though, and this is perhaps the killer: he found a further sliver showing the edge of the present f. 102, and it reveals the edges of what we know must have been on the lost f. 107, namely the continuation of Du Fay's Iuvenis qui puellam. That makes it clear that the middle of the gathering was at ff. 104–5. If the manuscript had regular ten-leaf gatherings and the foliation began only on the first full opening, then it would have the traditional (for Italy) flesh-side on the outside of each gathering.

That the portions of twelve leaves that we now have are all on parchment certainly suggests that the entire parent manuscript was on parchment, which is not the case for any of the surviving complete Italian choirbooks of those years. The copying is far more consistent and controlled than those other manuscripts, and the use of red ink more generous than in the others. There is a good chance that it indeed had regular ten-leaf gatherings throughout (even if f. 22 is placed backwards).

Now here's the beautiful bit. Klugseder found no more fragments, but he did find five more former Weihenstephan books with binding-boards that have ink-stains of pages that plainly came from the same manuscript. All that needs to be done is reverse the mirror-images and, with modern photographic techniques, it is possible to read just enough to identify some of the music (if you happen to be Margaret Bent). So those images are presented here alongside the relevant bars from Bent's own recent facsimile of the Bologna MS Q15 – itself endorsing the guesses of north Italian origin some time around the 1430s.

What is entirely frustrating here is of course that those further leaves are (or at least were) lying around somewhere in the Bavarian State Library, because if the binding-fragments of two books were kept complete on restoration, there is no reason to think that those from the others were discarded. Even worse: Klugseder demonstrates that each binding used four leaves of the original manuscript, two overlapping at the front and two overlapping at the back. So those five new books probably contained bits of a further twenty leaves from the manuscript MuL. And the frustrations do not end there: books printed up to the magical date of 1500 are very well catalogued and it is easy enough to see, worldwide, where other books formerly at Weihenstephan may be; for books after 1500 there is no such help. Moreover, the two new bifolia in Vienna – the ones that started this entire project – were used quite differently from those at Weihenstephan: the bifolia were kept intact and it is most likely that they were used already in Venice as provisional bindings for books that were intended to be bound more permanently at a later date. The book from which they come was printed in 1516, so perhaps all the bindings concerned are from the second decade of the sixteenth century. (Klugseder briefly argues otherwise but seems not very convinced; Bent seems even less convinced.) After all, that is roughly when you would expect a manuscript of polyphony from the 1430s to be split up and used for binding.

The present luxurious publication offers full-size colour plates of all the twelve fragments except the – to my mind – crucial slivers from ff. 22 and 102, which are in black-and-white only, buried in the German part of the book on pp. 33–4, where it took me an age to locate them. In addition, it has Margaret Bent's editions of all the unpublished music, and extensive commentaries by the two authors (Klugseder on the physical, Bent on the musical) in both German and English.

One sadness is that the publishers have made no attempt to align the fragments roughly where they would have been in relation to one another on the original pages: everything is brought up to the same upper margin, even when it is just the bottom portion of the sheet. That's an understandable decision, perhaps: to do otherwise would have entailed printing on paper larger than the standard 30 cm and would have increased the already substantial price.

Another sadness is a tiny point but worth nailing here and now. When Isabelle Ragnard first noticed that much of the related manuscript in the Bologna University Library (MS 2216, known to its friends as ‘BU’) had staves ruled with a double rastrum, namely two staves at once, we all started thinking afresh about these things.Footnote ¹ But the situation there was quite clear: ninety-six pages all ruled exactly the same way, with the staves alternating 18 and 16 mm and with a couple of examples of wriggles to prove that it really was a double rastrum. The situation with MuL is entirely different. BU is on paper, which remains relatively stable; MuL is on parchment, which curls and expands (and it is easy to see from these lovely photographs how distorted the thin parchment has become). Klugseder's view (pp. 41–3) that the staves alternate between 12.5 and 13 mm simply cannot be endorsed: half a millimetre can happen simply by holding the rastrum at a different angle not to mention the way parchment can warp and bend over six hundred years. That he has those distances alternating on some leaves but not on others is in itself a demonstration that we are not dealing with a double rastrum. And there is no sense in any notion that the stave-ruler had two different rastra which he used one after the other more or less at random.

But the book stands as a clear statement that fragments of this kind are of major importance for music historians and that librarians must be encouraged to make them known. That way, perhaps we will get even more of what looks to have been one of the seriously authoritative sources of its generation.