Allegri’s Miserere:
Manuscript Sources


This research has largely been superceded by the much more detailed work in Graham O'Reilly's book Allegri's Miserere in the Sistine Chapel (Boydell Press 2020). I recommend it to anyone interested in the subject.

Mss Capp. Sist. 205 & 206

These are companion volumes of different settings of the Miserere by a number of composers, including Palestrina, Anerio, Nanino and several anonymous versions. and is dated by a mention of the reigning pontiff, Alexander VII, (1655 - 1667). MS 205 gives the year more precisely as 1661, nine years after Allegri's death. It is thought that the work was conceived in 1638. The pattern of performing this work with alternating choirs had already been set, as MS 205 contains verses 1, 5, 9, 13 and 17 of each work, and MS 206 contains verses 3, 7, 11, 15, and 19. The last verse, 20, is contained in both.

The work is a falsobordone, alternating between a five-part and four-part choir, each separated by plainsong. The original part requirement, taken from the clefs, is SATTB and SSAB, and would be performed one voice to each part. The clefs give no indication of transposition, and the work is essentially homophonic, with some rhythmic interest. The first chord of each verse is speech rhythm, sung to as many words as necessary. The first choir (a 5) contains a number of suspensions which lead towards the final cadence, as does the second choir (a 4), whose other main feature is the striking 4th chord, with a diminished 5th between the upper parts, eventually resolving into D Major. The second choir contains no top C, but merely a slow descent from a treble E flat to the final cadence. The last verse has all nine voices singing together. The work has always been associated with Holy Week, and Psalm 51 features in the Liturgy for Lauds on Maundy Thursday and Good Friday, commonly called Tenebrae (darkness) from Luke's Gospel: "and darkness covered the whole earth".

An edition of Allegri's original falsobordone, without ornamentation, based on these manuscripts, is available for download. Other editions of the work can also be purchased.

MS. Capp. Sist. 185

Almost one hundred years after the conception of Allegri's original, this manuscript appears in a volume dated 1731, scribed by Johannes Dominic de Biondini (1725 - 1750). Three other people are also mentioned; Clement XII, Pope from 1730 to 1740; Cardinal Otthoboni, Vice Chancellor of the Bishop of Tuscany; and Ansano Bernini, Master of the Papal Chapel. The work is clearly marked with Allegri's name on the first page of music, yet there are a number of differences from the earlier sources. The first tenor part has become another treble part, and some of the bass notes are changed to provide a more effective harmony. There is also more independent movement in the harmony, and some ornamentation, including a curious C# at one point. It is here that the marvellous chord of C minor with a minor 7th and 9th is introduced, before the cascade of treble thirds. This matches discordant moment in the original second choir. The work also takes on a much more polyphonic character, but the speech rhythm is retained. Allegri's original part requirement returns for the last verse. The style of print is relatively archaic, and the music is written out in parts, as a choir book.

Although no matching parts exist for the second choir, Biondini wrote a complete set, MSS 340 & 341, in 1748. These contain the choirs a 4 and a 5 respectively, of both Bai's and Allegri's settings. The 5-part choir in ms 341 matches that found in MS 185. MS. 263 is an autograph collection of compositions by Johannes Biordi (1691 - 1748), Master of the Papal Chapel from 1737 until 1742, which also includes Allegri's Miserere for 5 voices, before his own setting of the text, which has two choirs of 4 voices before a final 8 voice last verse. A version is also found in MS Capp. Sist. 354, a book of Masses, Motets, Hymns and Psalms of 1705, which uses Allegri's original music, but switches the text, so that the 4-part choir sings the first verse (and so on).Up to Top

The British Library Sources

Charles Burney (1726 - 1814)

La musica che si canta annualmente nelle Funzioni della Settimana Santa nella Cappella Pontificale. 1771
This printed book is relatively important, since it represents a different source from the other mss. The first choir is a further reworking of MS 185, 40 years later. However, the harmony is slightly more structured than before. The legend Si Canto il Mercoledi e Venerdi Santo Miserere del Signor Gregorio Allegri is visible on the first page of music, showing that it was sung on Wednesday and Friday of Holy Week. Each verse of each choir is harmonically identical, whereas MS 185 has a number of differences. An attempt to notate the speech rhythm is made only in the first verse of each choir; after that, subsequent verses give the indication Canto Fermo above the speech rhythm.The second choir shows Allegri's original, with some slight alterations, removing some suspensions and the striking 4th chord. In the last verse we find the instruction Questo ultimo verso si canta Adagio e Piano, smorzando a poco a poco l'Armonia. The book also contains Bai's setting of the same text, sung on Maundy Thursday, which is remarkably similar to other eighteenth century sources, and Stabat Mater, Improperia (Reproaches) and Fratres ego by Palestrina, which were sung as a psalm in Mass, at the Adoration of the Cross, and on Tuesday of Holy Week respectively.

British Library Add. 31395

Bought by me at Rome in the Corso, 1841. Frederick Blaydes, Ch. Ch. Oxon.
This copy is all but identical to Vatican MSS 185 and 206, but scored and collated. The only difference is that the speech rhythm has been written out in a very deliberate style in all but a few verses. Interestingly enough, the speech rhythm here is consistent with that present elsewhere. Allegri's authorship is clearly stated.

British Library Add. 24291

Musica Clasica (sic) by several Maestri della Cappella Vaticana, copied by or for Giovanni Jubilli.
This scored MS starts off as before, with the first verse a copy of Vatican MS 185, and the choir a 4 is in keeping with Allegri's original. However, the copy is not complete; he allows one more first choir verse, and then a quite different second choir verse before the last 9 part verse. This verse a 4 is verse eleven of Tommaso Bai's setting of the Miserere mei. An organ part is written out for each verse, being a simple reductio partiturae.

British Library Add. 2468

The structure of this Miserere is the same as that of Allegri's, being in the same key & with a similar modulation, although not so much filled internally, and is probably the incorrect copy mentioned by Hawkins, vol. 4 p. 90. See the original by Burney.
The above remark is contained in the British Library catalogue. Indeed, it is unquestionably a similar work to Allegri's, but rather than being a wrong copy, I would suggest that it is a deliberate translation for the needs of a less resourceful choir. The piece is now 4 part throughout, being scored SATB in every verse. The music has also reverted to a more chordal nature. The reference to Hawkins relates to a comment made in A General History of the Science and Practice of Music, published by Sir John Hawkins (1719 - 1789) in 1776. In it, he mentions that most copies of the Allegri are incorrect, but I take this to mean that they lack the abbellimenti.

British Library add. 2470

Collection of Compositions written for the church of Rome in the same hand.
This MS is a 19th century copy belonging to Rev. Edward Goddard. Surprisingly for so late a copy, it is identical, save for a slight change in underlay to the Vatican MSS 185 and 206, scored and collated. It is significant that this version should still be punted about in the 19th century, at a time when the more contrived abbellimenti were all the rage. (See Other Sources.) Up to Top

Ornamented Sources

British Library Add. 31525

This scored source shows an 18th century MS of Bai's setting, copied by Benedetto Morganti, brought back from Rome by Joseph Warren, although the catalogue notes that it is different from Bai's setting as reported by Burney. A swift glance shows it to be infinitely more than that. It is clearly not Bai's setting, but Allegri's. The music resembles MSS Capp. Sist. 185 and 206 as all other 18th century sources have done. However, both choirs are highly ornamented with trills, runs, and other adornments, and there are also some extreme dynamic indications. In the first choir, they elaborate on the movement already present in MS 185, and in the second choir, they run amongst the held notes in the two treble parts, each taking his trill in turn.

The ornaments, are clear in their effect and intention, if a little uncertain in their notation. However, the first half of this choir clearly mirrors the top C version, which is taken from Alfieri's 1840 notation. The slentando indication in this first half is matched in Alfieri by an elongation of the lower parts to accommodate the first treble's run, which include a C# in both, echoing the similar ornament in the first choir. Each verse of each choir displays identical ornaments. However, the last verse is totally unadorned, and is identical to Allegri's original.

On the next folio, a second version follows which is a similar version of the same thing. The ornaments are very similar, and in all the same places, with a few slight variations, but the final verse has been adapted, with only one tenor part. This version is sloppier calligraphically but clearer musically.

Pietro Alfieri (1801 - 1863)

In 1840, Alessandro Geminiani, a nom de plume for Pietro Alfieri, published Il Salmo Miserere posto in musica da Gregorio Allegri e da Tommaso Bai, Publicato cogli Abbellimenti per la prima volta. (The Psalm Miserere, put in music by Gregorio Allegri and by Tommaso Bai, published with the ornaments for the first time.)

A priest in Rome, he taught Gregorian chant, and consistently tried to save the musical heritage of the Vatican from modern influences. The failure of this task, in his eyes, and the Church's refusal to publish his plainsong books, made him lose his mind towards the end of his life. However, this volume should be thought of as a successful attempt to save the past. A lengthy introduction announces that these works are two of the most beautiful sacred compositions ever written, and as such, should be made public, especially because of 'inexact copies...without explanations'. Alfieri believes himself to be 'doing a kindness to lovers of sacred music'.

He further informs that the works are presented unadorned, with the abbellimenti in Appendices A, B, C, D & E, these letters being also placed in the score at the relevant points. Breves, when at the starts of verses, should not be taken literally, but sung to the nature of the words, and he also tells us that the Papal Choir 'convert G into B diapason'. I take this to mean that the works are transposed up a third, supported by Mendelssohn's transcription of 1831, which suggests that they sang up a fourth! In recent years, he confides, they have sung mixtures of both works, as follows. For both works, Bai's first verse was followed by Allegri's Amplius (verse 3). The other verses were sung as intended in each, although Bai's 8 part tunc imponent (the last verse) follows that of Allegri's in Allegri's version! This mixing of the two works must be the cause of the confusion between the two works and the 19th century scores that are a 'hybrid' of both works.

Allegri's version is presented straightforwardly, and is, again, as in Vatican MSS 185 and 206. The first choir is not ornamented at all, and the second choir ornaments contains similar elements to those found in MS 31525. Appendices A & C give the ornaments for the first half of the second choir verse in both Allegri's and Bai's version; B & D are similar endings for the second half of the verse in Allegri. E is an ornament for Bai's setting, for the end of the last second choir verse (Benigne fac..). All the ornaments are essentially similar, and all follow those in MS 31525. The main difficulty is that the letters placed over the music are not always the correct ones.

An edition is available of the work, based on Vatican MSS 185 for the first choir, and MS 206, with the ornaments as shown in Alfieri's record and BL Additional MS 31525, showing how the work would have been sung in the late 17th and early 18th century. Click here to purchase.

Julius Amann

Julius Amann's commendable study, Allegris Miserere und die Auffuhrungspraxis in der Sixtina, (1935) is a detailed record of all sources of Allegri's Miserere, variations in ornamentation and notation of speech rhythm. Despite this record, he only produced one edition of the work, a rendering of Allegri's original composition from MSS Capp. Sist. 205 & 206, with note values halved. In contrast to his meticulous research, there are a few errors in both the editorial notes and incipits. The book gives a German source for its record of the ornamentation, which is almost identical to that found in MS 31525. Up to Top

The 'Top C' Version

The received version, as it is widely held today, is a mix of Burney's first choir with a bizarre second choir, congealed into life in the first edition of Grove's Dictionary of Music & Musicians in 1880. In 1932, Robert Haas took Burney's first choir and final verse, adding W.S. Rockstro's second choir from Grove's. This consists of Burney's small choir with Alfieri's embellishment recorded in 1840 for the first half; and Mendelssohn's similar record of 1831 for the second half. The problem is that the Mendelssohn embellishment is a record of the first half, which was apparently sung a fourth higher than written at the time of his visit. It is this that causes musicologists to squirm with the bass jumping from an F# up to a C, followed by the swift gear change into C minor. Ivor Atkins, mirrored Haas's work for his English language edition of 1951, and also chose the plainsong for his edition somewhat spuriously.


The result is strangely beautiful, and probably here to stay. It is, after all, one of the most popular pieces of sacred music. However, it is neither a representation of the performance practice of the Sistine Chapel choir, nor a true reflection of how the piece was ever sung there.

An edition of the Miserere, tracing the evolution of the work from simple falsobordone to the erroneous Top C, can be downloaded here.Up to Top

Other Sources

There are a great many manuscripts of the Miserere in libraries across Europe; perhaps that is half the problem.

There are a number of Italian sources, which stem from outside the Vatican, dating from c. 1820 onwards that are completely removed from anything inside the Vatican, and owe more to contemporary influences. Whilst ranging in style, from the modest to the hugely operatic, they all claim to contain the famous abbellimenti. However, these versions are so elaborate as to render them bastard sons, with no claim to their father's title. Some are orchestrated, some contain many more parts, and nearly all have a musical language that can only belong to the 19th century. The waning of the Papal Choir during the 19th century casts further doubt on such sources, particularly whilst earlier, and more truthful sources abound. Up to Top


Whilst the Vatican sources make no reference to the plainsong verses, all of the British Library mss indicate the plainsong only by the first few words of each verse, without any clue as to the chant. MS 31525 states Il Populo risponde altro verso, whilst other mss variously mention that they are to be sung sotto voce, and also in coro, which suggests that the choir was responsible for the plainsong verses.

The work is based on the Tonus Peregrinus, or Wandering Tone, and so would have originally alternated between this chant and the harmony. However, over the span of time through which this piece was performed, a great many chants will have been used, including those found in the liber usualis in the liturgy for Lauds on Maundy Thursday and Good Friday (VIII 1 and VII 2, respectively). It is also thought that in the nineteenth century, the plainchant was merely recited on a single note. Up to Top

Tommaso Bai (1650 - 1718)

Capp. Sist. MS 203 contains Tommaso Bai's setting, unadorned, shortly after its composition, and was scribed by Thomas Altavilla in 1713, who cites Hieronimo Bigello as Master of the Papal Chapel at that time. Bai's setting of the same text is based upon Allegri's, being harmonically similar in its structure. However, there is more variation and harmonic elaboration, with each verse showing differences. The piece is written for the same voices as Allegri's original, SATTB & SSAB, but is more 'modern' than Allegri's slightly archaic setting. All speech rhythm has been removed, and the work takes a more chordal nature with underlying polyphonic movement. The speech rhythm is consistent with that used later for Allegri's. The last half verse, tunc imponent, becomes an antiphonal double choir of SSAATTBB, but again holds Allegri's original structure within it. This setting is, however, not interchangeable with Allegri's original nor any of the subsequent versions.

It is interesting that Alfieri informs us that the same abbelimenti are used in both Allegri's and Bai's versions. This again leads us to the conclusion that the performance is not ad lib, as it usually thought. These were set-piece cadences that existed before the music was composed!

It is also worth mentioning that Bai's first verse is different still from the 18th century 're-writing' of Allegri's five-part choir.

Bai's setting of the Miserere, with the abbellimenti, is available for download. It is set for two choirs of SATTB & SSAB, with an eight-voice double-choir last verse. Up to Top

The 19th Century and beyond

The latest Vatican source is MS. Capp. Sist. 375, dated 1892, and was written by the then Master of the Papal Chapel, Dominico Mustafa (1829 - 1912) who was himself a castrato. The catalogue describes it as Bai et Allegri Psalmus 50; ff 1 - 10v, and indeed, it does appear to be a hybrid of the two - mostly Bai's version, with occasional second choir verses taken from Allegri. The setting is ornamented, and Mustafa urges the Pope not to let anyone outside the Cappella Sistina Choir view the work. Mustafa's predecessor, Giuseppe Baini (1775 - 1844) had composed his own ten-part setting of the Miserere, which supplanted the Allegri/Bai hybrid for a time.

I suspect the reason for the mixing of the two works is as follows: most of Bai's four-part verses contain the abbellimenti, except the first one, Amplius lava me; so it makes sense for Allegri's setting of that verse to be transplanted. Similarly Bai's setting does not contain music for the first half of the last verse, Tunc acceptabis sacrificium iustitiae, oblationes et holocausta. Because Allegri does set music to this line, Allegri's setting of that text finds its way into Bai's version. This roughly fits the mixing of the works as described by Alfieri.

Mustafa had composed his own six-part setting of the Miserere, and as a castrato had sung the Miserere himself. As "director in perpetuity" of the Cappella Musicale, he conducted its performance with 'the last castrato' Alessandro Moreschi (1858 - 1922), whose voice could cope with the rigours of the work. The work was now performed up a fourth from written pitch, and so still required a top C, but in the first half of the four-part choir, rather than the second half. In 1903, Pope Pius X banned castratos from the Cappella Musicale Pontificia, and boys were reinstated.

For most of the 20th century, choirs have sung and audiences have listened to a version of the Miserere that bears little relationship to anything that was sung in the Sistine Chapel.Up to Top

<<| Introduction | Conclusions & Theories  >>