John Michael Cooper (University of North Texas)

“. . . da ich dies Stück gern recht correct erscheinen sähe”: Philological and Textual Issues in Mendelssohn's Hebrides Overture, Op. 26

[Introduction] 

I. Genesis and Publication History of the Overture: A Chronological Sketch

II. The Musical Sources

III. The Revision History

IV. Conclusions

[Printable version]

One of the first of many remarkable moments in Mendelssohn’s Hebrides Overture occurs in mm. 7-8. Here, after the evocative opening pitting strings and woodwinds in sustained open octaves against sequential statements of the gapped scale that comprises the Overture’s main motive, the first bassoon, violas, and cellos offer a darkly colored surge, perhaps evocative of the swelling and breaking of a wave on the shores of Scotland’s fabled outer isle (Example 1a). This picturesque gesture dates from early on in the Overture’s history: it is present in all surviving autograph orchestral scores, three surviving contemporary copies of the full score, the composer’s own arrangement for piano duet, and the printed parts issued by Breitkopf & Härtel in March 1834. It also is retained in the edition presented in Series 2 of the Breitkopf & Härtel series of Mendelssohn’s Collected Works issued in the 1870s,[1] as well as most modern editions.

It may come as a surprise, then, that Mendelssohn abandoned this gesture in his latest documented revisions of the Overture. The first edition of the orchestral score, as corrected by the composer and released more than a year after the publication of the orchestral parts, replaces the familiar swell with another thematic element, the signature gesture of the consequent phrase of first subject itself – a gesture that in the familiar version is absent from the introduction (Example 1b). Nor is the first edition of the orchestral score the only documentation of the proliferation of this late revision, for at least one set of manuscript parts from the mid-nineteenth century transmits the same variant. The result is that this early memorable moment in the work is called into question not by unreliable later editions (as is usually the case), but by Mendelssohn himself, as well as surviving records of mid-nineteenth-century practice.

Despite recent decades’ progress in understanding the composition of the Hebrides Overture and the abundant changes introduced during its genesis, significant questions remain. Many of these questions concern the philological relationships among the manuscript and print sources and their respective degrees of textual authority. This study addresses some of these questions in order to construct a firm chronological framework for clarifying the Overture’s more difficult textual issues. A review of the composer’s correspondence and a correlation of the large- and small-scale revisions reflected in the surviving witnesses permits a detailed reconstruction of the successive stages in the textual development of this Mendelssohnian masterpiece – the stations along the way in its journey from an evocative sketch to one of Mendelssohn’s most widely acclaimed instrumental works. The undertaking also offers lessons concerning the issues that confront all who wish to know Mendelssohn’s music through reliable musical texts.

.

I. Genesis and Publication History of the Overture: A Chronological Sketch

The genesis of the Hebrides Overture has been thoroughly documented in a number of previous studies, especially those by R. Larry Todd.[2] The following pages summarize the essentials of this genesis with an eye to the events and documents that pertain directly to the work’s source-situation. The analytical and interpretive insights offered by the relationship between the sketches and the printed score, already discussed in detail by Professor Todd, are beyond the scope of the present study.[3] 

The documented origins of the Overture represent a striking convergence of biographical, poetic, and visual impulses. The work was conceived during Mendelssohn’s trip to Scotland in August 1829. The date of 7 August was particularly illustrious: on that date the composer recorded his impression of one of the scenes he encountered in a finely wrought pencil drawing[4] and began a letter to his family that not only describes his experiences verbally, but also musically responds to that scenario in a short-score sketch of the work’s opening twenty-one measures (see Figure 1).[5] Although Mendelssohn frequently referred to the work in letters to his family, he was able to complete a full score only fifteen months later, while in Rome. A set of sketches may be securely dated as stemming from mid-October 1830,[6] and a letter to his family dated 30 November 1830 records that he intended to send a copy of the score to his close friend Eduard Rietz as soon as it was finished.[7] In a letter to his father dated 10 December 1830, he expressed his wish that the Overture could be finished by the next day, his father’s birthday.[8] There is, however, no evidence that this wish was fulfilled.[9] The earliest surviving complete autograph (identified as source AR later in these remarks) bears the closing date "Rom d. 16 Dec. 1830."

Once he had completed this version, Mendelssohn began to share his new masterpiece with others: he reportedly played the work for Hector Berlioz during the French composer’s time in Rome (probably during the two composers’ first encounter, in March 1831).[10] In the meantime, he had dispatched a manuscript copy to his older sister, Fanny, with useful details about the work’s state at that point (see Figure 2):[11] 

The evolving work’s informal dissemination continued. Despite efforts to get his orchestral music performed in Paris, he reported to Fanny on 21 January 1832 that he refused to release the Overture there because he "still [did] not regard it as finished";[14] nevertheless, he did show the "provisional score" (provisorische Partitur) to his close friend Ferdinand Hiller sometime during the Parisian sojourn, and in London he played it for Ignaz Moscheles on 30 April 1832.[15]  

On 5 May 1832 Mendelssohn reported to his family that the problematical "middle section" in D major had been eradicated and he was writing out the entire score anew in preparation for an upcoming concert by the Philharmonic Society of London; despite apparent hopes that the work be included in the Society’s concert on 7 May, the premiere was delayed because at the time of the rehearsal on 5 May the score was "not yet written out." Nevertheless, the composer considered that he had by then "made The Hebrides significantly different and better."[16] His diary notes that he did complete the new score on 6 May,[17] and on that same day he gave a score of the version originally completed on 16 December 1830 to Ignaz Moscheles. Of this remarkable gift, Moscheles later recounted that he had contested the need for the revisions undertaken in the meantime but was unable to dissuade the composer from his "principle of revision."[18]  

The new version was rehearsed on 12 May and premiered as the Overture to the Isles of Fingal at a concert given by the Philharmonic Society of London under the direction of the composer’s friend Thomas Attwood on 14 May.[19] A second performance was given on 1 June by the same ensemble, now under the direction of Sir George Smart. Mendelssohn gave "the score of [his] Overture to the Isles of Fingal" to the Society at a reception held on 6 June 1832.[20] Now pleased with the work, he began to take concrete steps toward its eventual publication. On 19 June he inscribed the date on an autograph arrangement for piano four-hands (see source APf, below) and made a corresponding entry in his diary,[21] and the following day he inscribed an orchestral full score (see source AL, below).

But if the Overture appeared to be moving rapidly toward completion in the summer of 1832, circumstances conspired to stall this progress. The geographic distance between London and Berlin had permitted the young composer to continue his productivity despite the professional issues that arose after the death of his composition teacher, Karl Friedrich Zelter, on 15 May 1832. But when he returned to Berlin he was flung squarely into the middle of the intense rivalry for the directorship of the Singakademie, a process that extended into January 1833 and ultimately resulted in an embarrassing defeat for Mendelssohn.[22] Despite these challenges, this period witnessed two further performances of the Hebrides Overture, on 10 January and 14 February 1833.[23] Both events were well received, albeit not without some reservations from the critics: Ludwig Rellstab suggested that "[p]erhaps the fault of the composition is only that it requires a commentary,"[24] and an anonymous reviewer in the Leipzig Allgemeine Musikalische Zeitung commented that "the Hebrides Overture by F. Mendelssohn-Bartholdy was too serious for this concert-going public and may also be less musically self-sufficient than the Midsummer Night’s Dream Overture."[25]  

The final steps toward publication evidently were taken in mid-1833. Mendelssohn initiated arrangements for the printing of the piano-duet version with the London firm of Mori & Lavenu, and (in accordance with publication conventions of the day) he proposed German publication of this version to the Leipzig firm of Breitkopf & Härtel in August; by then, the London firm had settled on a release date of 1 October.[26] On 18 September, traveling through Leipzig on his way from Berlin to Düsseldorf (where he was to assume his new position as Municipal Music Director), he wrote again to Breitkopf & Härtel to remind them of the need to advertise this date, also specifying the title for the German editions of both the orchestral version and the piano-duet arrangement.[27] That schedule proved untenable, however, for as of 27 September the Leipzig firm had heard nothing from its London counterpart.[28] On 4 October Breitkopf & Härtel notified the composer that the piano-duet arrangement had been received from Mori & Lavenu, also asking that he send the written-out orchestral parts as soon as possible.[29] Finally, the piano-duet arrangement appeared on 15 October 1833 (see discussion of source EPf1 below).

Further complications arose in the production of the orchestral parts. In response to a letter of 21 November 1833 from Breitkopf & Härtel asking that the manuscript parts used for the Berlin performances be sent to them, the composer’s mother, Lea, forwarded this letter to the composer with a note that "Fanny says that the parts are not here."[30] Mendelssohn reported this problem to the publisher on 29 November, conjecturing that the parts had been lost. With this letter, however, he sent a corrected full orchestral score for the work and asked that the parts be engraved from it, with particular attention to the placement of dynamics and expressive indications. He also stated that this score was provided with rehearsal letters that were to be copied into the parts, and specified that the published score be grouped with the Midsummer Night’s Dream and Calm Sea and Prosperous Voyage Overtures under a single opus number:

Today I am sending via express post the score to my Hebrides Overture, and I regret to have learned from Berlin that it was not possible to send you the written-out parts, as you had requested. They believed there that I had the parts here with me, so I fear that the parts have been lost. In any case, I ask that you have the parts made from this score, in which I have changed a few more things and placed all indications very precisely; also, please ask the engraver to be very exact in the placement of p, f , crescendo, etc., and of the rehearsal letters, which must be in all the parts. The cello and bass need not always be written in two staves, as occurs in the score, but only when the cello part diverges from the bass part. I would very much appreciate it if you could send me proofs of these parts, for I very much want this piece to appear properly.

At the same time, permit me an inquiry: in your Allgemeine musikalische Zeitung I was sharply scolded for not having released the score for my Midsummer Night’s Dream Overture; it was considered my fault and attributed to a fear of criticism. Quite the contrary, however, it has long been one of my dearest wishes that some of those scores that I myself prize most should also be presented to the public, for I believe that they would do no disservice to my name. Until now I considered this impossible, and even now I would not ask you to produce such a publication if not for the above-mentioned review, and if other considerations did not lead me to believe that such a publication would perhaps be feasible. I therefore wish to ask whether you might publish the Midsummer Night’s Dream Overture, the Hebrides Overture, and a third one of the same sort in score? They would have to be assigned a single opus number and would require far fewer pages than a Beethoven symphony, so they could become fairly profitable. In such a venture I would naturally relinquish my honorarium and would be very happy simply to see my wishes met. Because you have published so many of my compositions one after another I would not have proposed this to you, but because you hold the rights to two of these overtures I had no choice but to ask you first ...

P. S. Please give the title of the Hebrides Overture in German, as I specified earlier.[31] French titles are bêtes noires for me.[32]

In a performance generally overlooked in commentaries to date, the Overture received its Leipzig premiere in a benefit concert for the poor of the city on 13 February 1834;[33] in response to this occasion an anonymous reviewer for the Leipzig Allgemeine Musikalische Zeitung referred to "the new, nicely wrought Hebrides Overture, published by Breitkopf & Härtel."[34] Breitkopf & Härtel issued a new announcement of the arrangement for piano-duet on 5 March, along with a notice that an arrangement for piano solo was at press.[35] On 14 March 1834 Mendelssohn wrote from Düsseldorf to send the publisher his specifications for the title pages of the score editions of all three overtures:

...You prefer to release the scores for the overtures individually and I understand that this must be much more advantageous for you. Nevertheless, I wish that a collective title for all three be made (perhaps something like "Three Concert Overtures: No. 1 Ein Sommernachtstraum; No. 2 Die Hebriden; No. 3 Meeresstille und glückliche Fahrt"), and that additionally each title page would read Three Concert Overtures: No. 1 Ein Sommernachtstraum; No. 2 Die Hebriden, [and so on] so that it will be clear that they belong together. This leads me to mention that one of the best painters we have here has offered to make a small vignette for each title page, which obviously would be very nice. I would very much appreciate it if you would agree to this, but in that case the entire set of title pages would have to be engraved here, and first of all I must ask you approximately what you pay for your title pages and what the specifications of your format are. Finally, I must unconditionally request that I have an opportunity to proofread the three scores before they are released. When do you expect to give them to the engravers? For various reasons I hope that this will occur soon, and ask once again that you answer my questions, especially the last one, soon, also telling me whether the orchestra parts for The Hebrides have been released.[36]

The parts for the Hebrides were published in late March 1834, and the publisher wrote to Mendelssohn on 21 April requesting that he provide the dedication by mid-May because the scores for all three overtures were to appear by the end of that month.[37] A second Leipzig performance was given on 5 May,[38] but once again delays ensued in the publication of the scores. The orchestral parts were reviewed (along with the piano-duet arrangement) in the Leipzig Allgemeine Musikalische Zeitung on 25 June,[39] and on 29 July the composer acknowledged receipt of these parts, also stating that their production left nothing to be desired.[40] The same letter accompanied the corrected proofs for the Midsummer Night’s Dream Overture and the Rondo brillant, Op. 29, as well as a corrected proof of the collective title page for the three overtures, with the dedication to the Crown Prince of Prussia written in by the composer. The letter also specifies that the three overtures were to be given the collective opus number 27 – a wish that remained unfulfilled, although that number was eventually assigned to Meeresstille und glückliche Fahrt – and conveyed Mendelssohn’s hopes that the proofs for the other two overtures in the triptych could soon be sent to him for revision. The publisher replied on 8 August that the engraving of the Hebrides and Meeresstille would soon be finished[41] and Mendelssohn acknowledged this response on 14 August.[42] On 15 November he returned the proofs with a lengthy cover letter that referred to several further necessary alterations:

...Enclosed please find the proofs of the two overtures. I have had to make still more changes here and there, and therefore ask you to request that the engraver follow all my comments exactly. In addition, it would be very good if the overtures could be released rather soon, also because of the dedication, for which I requested and received permission long ago.[43]

Another Leipzig performance of the Overture, now billed as Ossian in Fingalshöhle, occurred on 4 December 1834, presumably using the parts published several months earlier, even though they would not have included the necessary changes he had entered in the score proofs returned on 15 November. The only other development in the last months of 1834 was the appearance of Friedrich Mockwitz’s arrangement of the Overture for piano solo – an arrangement in which the composer had no hand, and to which he objected strenuously.[44]

On 16 January 1835 Mendelssohn wrote from Düsseldorf that he had received no response to a letter (now lost) of 4 December, pressed for an update, and asked that presentation exemplars be sent directly to his father in Berlin so that he could forward them to the Crown Prince as dedication copies.[45] The publisher evidently replied on 22 January, commissioning Mendelssohn himself to prepare an arrangement of the Hebrides for piano solo, but the composer declined because he felt that the work did not permit reduction for that medium.[46] On 28 February he responded happily to the publisher’s report (now lost) that the engraving of the scores of the three overtures had been completed, also asking that the separate edition of the Hebrides be dedicated to Franz Hauser (to whom the piano-duet arrangement had likewise been dedicated).[47]

Finally, the scores of the three overtures appeared in mid-March 1835.[48] In accordance with the composer’s specifications, the collective title page identified the Crown Prince of Prussia as the dedicatee; however, no separate title pages were provided for the individual publications. The composer’s acknowledgment of receipt, dated 10 April 1835, manifests both gratitude and relief.[49]

The genesis of the definitive version of the Hebrides Overture spans five and a half years, from the initial visual/musical impulse in August 1829 to the publication of the score in March 1835. The chronographic evidence is riddled with lacunae, ambivalences, and contradictions: some letters are lost, others appear have crossed in the mail, and still others document problems and points of confusion. Even taken by itself, this evidence poses challenges for the scholar who wishes to establish a reliable chronological framework for the successive versions of the Overture’s Notentext. But the surviving contemporary musical sources prove equally vexing, for Mendelssohn’s correspondence makes clear that the latest revisions date from November 1834 – well after the publication of the orchestral parts and the arrangement for piano duet. As will be shown in the following section, however, the difficulties posed by those sources ultimately emerge not as hindrances, but as keys – indicia that permit a remarkably precise reconstruction of the work’s textual history.

.

II. The Musical Sources

The musical source-situation for the Hebrides Overture is typical of Mendelssohn: the work’s protracted gestation is documented through his letters and a number of chronologically intertwined and geographically dispersed manuscripts. Some of these manuscripts are entirely or partially in Mendelssohn’s hand, others are significant contemporary copies, and still others are missing or lost. All offer significant variants from the published version of the work. The following inventory differentiates between autographs (sources wholly in Mendelssohn’s handwriting), contemporary manuscripts principally or entirely in another hand, and significant print sources. For convenience, brief identifications of these sources are provided in Table 1.

Autograph Sources

AS1. New York Public Library, *MNY: Sketch of mm. 1-21 in short score (see Figure 1, above). This sketch is contained in a letter from Mendelssohn to his family begun on 7 August 1829 in the fishing village of Tobermory, on the northeast side of the island of Mull in Scotland. Continued on 11 August, the letter bears the postmark 12 August. This passage is transcribed and discussed in detail in R. Larry Todd’s monograph on Mendelssohn’s major concert overtures.[50]

AS2. Bodleian Library, Oxford, MS M. Deneke Mendelssohn c. 47, fol. 29r. These undated sketches for the Overture’s recapitulation and coda are contained in a posthumously assembled volume of miscellaneous compositions by Mendelssohn, J. S. Bach, and Fanny Hensel.[51] On the basis of philological and biographical considerations, they may be securely dated as stemming from mid-October 1830.[52]

AR. Pierpont Morgan Library, New York City, Robert Owen Lehman Deposit.[53] This manuscript is the earliest surviving autograph full score for the Hebrides Overture. At the end of the score, on p. [42], is the autograph inscription "Rom d. 16 Dec. / 1830." Not surprisingly (given its chronological distance from the Overture’s completion), this version of the work exhibits many differences from the final one. It originally comprised 320 measures but was eventually reduced to 311 by cuts; this stands in contrast to the 268 bars of the familiar version. Some of these differences have already been discussed in the literature concerning the Overture, but they have yet to be explored fully in the context of a study specifically devoted to textual and philological issues.

In the center at the top of the first page of AR appears the title Die Hebriden; in the upper right-hand corner is the invocation "L. e. g. G." ("Lass es gelingen, Gott," an invocation that Mendelssohn frequently inscribed on his manuscripts at the beginning of a compositional undertaking). Above the first system of the score is the subtitle "Ouvertüre" (see Fig. 3). Although the pencil markings on the first page suggest that the manuscript was in the possession of Charles Gounod (1818-93) at some point, when he may have received it is unclear. It was purchased in 1911 by the Musikhistorisches Museum of Wilhelm Heyer (1849-1913) and auctioned again in 1926; at that point it was purchased by Hugo von Mendelssohn Bartholdy (1894-1975; great-grandson of the composer). A facsimile was published by Amerbach (Basel) in 1947. The manuscript subsequently was acquired by Robert Owen Lehman, who placed it on deposit in the Morgan Library in 1972.

Source AR comprises forty-four oblong quarto pages containing fourteen printed staves per side and measuring 19.5 x 23.5 cm; of these forty-four pages, the last two are blank except for the printed staves. Although the pages presumably were originally arranged in bifolia, at some point these were separated; each leaf is now tipped into the binding, which dates from 1960s Paris. The pagination, evidently in Mendelssohn’s own handwriting, appears on both recto and verso sides on pp. 2 through 19; thereafter, page numbers appear only on recto sides through p. 41. The final leaf of the score contains no page number, and the final recto and verso sides (pp. [43-44]) transmit neither music nor page numbers.

The paper of AR bears the watermark SWIDNITZ underneath a shield with posthorn on one side, and on the center of the other side F RITSCHEL. This watermark closely resembles that identified by Georg Eineder as no. 1562, which dates from 1830.[54] The foliation structure suggests that the manuscript originally fell into two halves, the first comprising pp. 1-20 and the second comprising pp. 21-[44]. As shown in Diagram 1, the first of these was a single gathering comprising five bifolia, but the second was more erratic; in particular, the fact that pp. 33-[44] originally comprised a single gathering of three bifolia but pp. 21-32 were a series of individual bifolia might suggest that the latter group posed some particular problem for the composer.[55]

Source AR also includes several extraneous markings. The most readily evident of these stem from the hand of Charles Gounod. As shown in Figure 3, the first page of the score includes a remarkable intervention by Gounod, who replaced the autograph whole-note rest in the contrabass line in m. 3 with a half-note D (without the concomitant half-rest) in fine blue pencil; next to the inserted note Gounod placed a "+" sign referring to an explanatory note at the foot of the page:"+ Je crois que le Ré a été oublié à la Contre Basse. / Ch. Gounod."[56] These original annotations were later supplemented (perhaps not by Gounod) with an "x" in fine red pencil, so that the original "+" signs appear as .

A second set of extraneous marks evidently stems from the hand of Ignaz Moscheles, to whom Mendelssohn reportedly gave the manuscript in early May 1832 (see above). These markings are in a coarse blue pencil. The first set frames an extended passage in which source AR varies substantially from the familiar version of the work: on p. 5 a large "X" is located above the system at the beginning of measure 6 (following m. 32 of the familiar version); a corresponding "X" (in the same script and written with the same implement) is placed beneath the system at the same point, and another large "X" (in the same hand and implement) is found beneath the system at the end of the passage (the end of the penultimate measure of p. 6, just before m. 47 of the familiar version). This passage is identified as No. 1 of the large-scale revisions discussed below.

Moscheles also placed a large "X" beneath the systems enclosing another passage that differs substantially from the familiar version of the Overture. The opening "X" falls beneath the system on p. 10, and the closing "X" is found beneath the system on p. 14; all totaled, this passage comprises thirty-eight measures that were eventually replaced by mm. 70-95 of the published score. As shown in Figure 4, beneath the second "X" Moscheles further noted, "Bis hieher X sind / 38 Takte verschieden von / den 24 gestochenen Takte" ("Up to this point X 38 measures are different from the 24 printed measures").[57] This inscription permits two conjectures concerning the dates of these annotations’ entry. Because it refers to "the printed measures" the second pair of Xs must date from after October 1833, the date of the first publication of the work; but if "the printed measures" were found in the printed orchestral score, the inscription must date from later than mid-Marchl 1835 (when that score finally appeared). Moreover, since the second inscription states that up to that point thirty-eight bars were different from their corresponding twenty-four [recte twenty-six] bars in the printed score and the second passage alone comprises thirty-eight bars, the first pair of Xs may have been written later on.

APf. Bodleian Library, Oxford, MS Horsley b.1.[58] This autograph is the composer’s own arrangement for piano duet (i.e., piano four-hands). As was customary for works for this medium, the score is laid out with the "Secondo" and "Primo" parts on facing verso and recto sides, respectively; each opening thus naturally gives the same measures for each player’s part. At the end of the Primo part (fol. 9r) the manuscript is dated in Mendelssohn’s hand: "London 19^th June 1832. / Für Marie und Sophie Horsley zu Erinnerung von / Felix Mendelssohn Bartholdy."[59] Despite its late date in the work’s compositional and publication histories, this arrangement still evinces some substantive differences from the familiar text.[60] These differences notwithstanding, the arrangement contains essentially the same 268 measures as the published version.

Source APf includes an autograph title page that reads: Overture / to the Isles of Fingal. / Arranged / as a Duet for two Performers / on the Pianoforte. / FMB. Above the first system of each part, the first page of the score itself is inscribed only "Overture."

Source APf is a part of the collection of Horsley family papers, acquired by the Bodleian Library in October 1987.[61] It comprises five sequential bifolia in oblong format, measuring 23.0 x 30.0 cm and containing fourteen printed staves per side; the folia with music inscribed on them (fols. 2-9) were joined together by side stitching, although the stitching is now partially disintegrated. These leaves are loosely enclosed in an outer wrapper (fols. 1 and 10) with an inscription that seems to be in the hand of Elizabeth Horsley: "Overture to the Isles of Fingal." The paper bears the watermark "W. King" on one side of the bifolia, and "1829" on the other. The Library’s cataloguer provided the folio numbers; there is no pagination or foliation in Mendelssohn’s hand. The autograph is unbound and kept in a modern envelope with other materials from the Horsley papers.

AL. Bodleian Library, Oxford, MS M. Deneke Mendelssohn d. 71.[62] This manuscript full score is the latest surviving autograph for Op. 26. It consists of a flyleaf followed by a regular series of bifolia, which are not grouped into any larger units but were pasted to one another at the edges, evidently before the manuscript was bound. The score begins on fol. 2 and concludes on fol. 18v. Following the last measure is the autograph inscription "London 20 Juni 32"; fol. 19r-v is blank. The flyleaf, in the hand of William Sterndale Bennett, reads: "William Sterndale Bennett. / Leipzig – March 17 – 1837 / N.B. – This Original Score was given to me by / Mendelssohn – WSB"; the leaf also bears Bennett’s stamp. The manuscript is in a dark green binding similar to that Mendelssohn used for many of his autographs. Since Mendelssohn did not inscribe the flyleaf, however, the manuscript probably was not yet bound when he gave it to Bennett; more likely is that Bennett had the binding done at some point later on and made the flyleaf inscription at that point. This autograph passed from Bennett to Prof. Thomas Chase, and from him to Mr. T. G. Odling. While in Odling’s possession, it was discussed in important articles by Ernest Walker and Gerald Abraham.[63] After these articles, however, the manuscript remained inaccessible to scholars for more than half a century. It was auctioned by Sotheby’s, London, in May 2002, and was purchased by the Bodleian Library. It now forms part of that library’s extensive collection of Mendelssohniana.[64]

The score of source AL transmits two paper-types, both in oblong format and measuring ca. 19.5 x 24.0 cm. Most of the manuscript (fols. 4r-17v) is on a nondescript paper bearing no watermark, but the paper of the outermost bifolia (fols. 2r-3v and 18r-19v) is browner, with darker stave ruling; this paper bears the same watermark as APf.[65] The musical content of these bifolia corresponds to that of mm. 1-27 and 258-68 of the published score, and in the distribution of measures over the pages these leaves correspond exactly to pp. 1-4 and 40-41 of the Rome autograph (AR).

Source AL originally contained 276 measures, eight of which were deleted via cross-hatching to yield the same 268 measures as the piano-duet arrangement (APf, EPf1) and published version (EP, ES, ER). In some places AL differs substantively from both of these versions. The first page of the score bears the main title Die Hebriden, and above the first measure is the subtitle Ouvertüre; the additional presence of the invocation "H. D. m."[66] concurs with the score’s extensive alterations and excisions to suggest that Mendelssohn regarded this manuscript as a new stage of revision (see Fig. 5). The score also contains a series of autograph rehearsal letters in pencil, evidently provided later on. The placement of these letters is as follows: A: fol. 4v (= m. 41); B: fol. 7r (= m. 77); C: fol. 10v (= m. 138); D: fol. 12r (= m. 165); E: fol. 13r (= m. 177); F: fol. 15v ( = m. 226).

AX. Autograph full score written in early May 1832 (lost). Mendelssohn’s correspondence reveals that he wrote out a new score for the Overture before the premiere (given by the Philharmonic Society of London on 14 May 1832) and that he presented a "score of [his] Overture to the Isles of Fingal" to the Society through George Smart on 6 June 1832; it seems reasonable to assume that these two were the same score. Such a score appears to have been listed in the catalog of the Society’s overture scores that was in use up to the 1840s, but disappears from the catalogs after that point. Nevertheless, as Peter Ward Jones has suggested, documentation of the Overture’s musical text as it was transmitted in AX appears to have survived in a manuscript copy evidently generated on the basis of parts produced from AX.[67] This copy is identified as source CL below.

Other manuscript sources

CO. Bodleian Library, Oxford, MS M. Deneke Mendelssohn d. 58. This apograph full score bears no date but does include a separate title page in Mendelssohn’s hand reading: "Ouvertüre / zur / einsamen Insel." Like the other manuscript sources, it is in oblong quarto format. There are also minor autograph corrections on fols. 4v and 8v. Bequeathed by Helena Deneke in 1973, the score was bound in the library.[68] On the whole, the musical content of CO concurs with that of source AR, but in CO the Overture runs to 316 measures rather than the 320 originally included in AR or the 311 to which it was eventually reduced. There are no rehearsal letters.

On the basis of calligraphic similarities between the dynamic and expressive markings present in CO and those explained in the composer’s letter to Fanny Hensel dated 25 February 1831 (see above), we may securely identify CO as the copy that Mendelssohn dispatched to Berlin on 26 February 1831 (cf. Figs. 2 and 6). This dating identifies ca. 25 February 1831 as either the terminus ante quem or terminus post quem of specific variants transmitted in AR.

CD. Sächsische Landesbibliothek, Dresden, Mus. 5543-N-507.[69] Like CO, This apograph full score is undated, and its musical content replicates that of source AR. It is a professional copyist’s manuscript in oblong format. It also contains a separate title page (in the same format) which reads, in one script: "Ouverture / (Hebrides, Fingalshöhle) / von / Felix Mendelssohn Bartholdy." Beneath this is another inscription in an unidentified hand: "Erste Niederschrift. / Später vielfach umgearbeitet und verändert" ("First draft. / Later extensively revised and altered"). The manuscript eventually passed into the possession of Wilhelm Wauers Musikstiftung (in Herrnhut, near Dresden). There the flyleaf was inscribed in 1878 "Copie der ersten Niederschrift Mendelssohns / 1. Titel von Rietz’s Hand. / W. W. 1878" ("Copy of Mendelssohn’s first draft / [The] 1^st title in Rietz’s handwriting"). The date provided in this inscription would suggest that Wauer acquired the score in the wake of the death of Julius Rietz (1812-77); however, neither the musical script nor the "first title" (presumably referring to the first four lines) appears actually to be in Rietz’s hand. Rather, the score is in the handwriting of Eduard Henschke (1805-54). Since Henschke was one of Mendelssohn’s preferred copyists in Leipzig – he was the copyist for the symphony-cantata Lobgesang and the E-minor Violin Concerto, among other works – the possibility cannot be ruled out that this copy was produced with the composer’s involvement.

Source CD contains the same 316 bars as CO. Unlike CO, however, CD does include rehearsal letters, added in pencil in a different hand. The placement of these is as follows (identified in terms of corresponding measures in AR and the published editions): A: AR p. 4, m. 3 [≈ m. 23]; B: AR p. 6, m. 6 [not in familiar version]; C: AR p. 11, m. 4 [not in familiar version]; D: AR p. 14, m. 3 [not in familiar version]; E: AR p. 16, m. 6 [≈ m. 112]; F: AR p. [20], m. 2 [not in familiar version]; G: AR p. 21, m. 8 [≈ m. 147]; H: AR p. 25, m. 2 [≈ m. 169]; J: AR p. 27, m. 2 [≈ m. 178]; K: AR p. 30, m. 1 [≈ m. 194]; L: AR p. 32, m. 5 [≈ m. 217]; M: AR p. 36, m. 1 [≈ m. 234]; N: AR p. 39, m. 5 [≈ m. 255].

A comparison of these rehearsal letters with those in AL and the first edition of the orchestral parts (see EP, below) suggests that the letters in CD were written quite independently of any input from the composer. This observation, together with the fact that CD was demonstrably copied from CO,[70] raises the possibility that CD bears little direct connection to Mendelssohn at all. Instead, its presence in the estate of Julius Rietz might suggest that it was a copy made for potential use in Rietz’s edition of Mendelssohn’s Werke. Such a scenario is plausible not least of all because by the 1870s the existence of a substantively different early version of the Overture was well known; indeed, both versions were performed in London in the context of the 1871-72 season of the Crystal Palace Concerts.[71]

CL. British Library, London, RPS MS. 114. This undated manuscript is another professional copyist’s manuscript, this time in the hand of William Goodwin.[72] It consists of a title page plus twenty-eight leaves measuring 23.5 x 29.5 cm. Unlike copies CO and CD, however, copy CL clearly relates most closely to autograph AL. As Peter Ward Jones has noted, there are some indications that this manuscript may have been generated from a set of orchestral parts extracted from the lost autograph written in early May 1832 and given to the Philharmonic Society by the composer on 6 June 1832 (see AX, above).[73] Of the surviving score copies, this is the only one that includes neither a programmatic title nor even an identification of the composer at the head of the score itself. Instead, this information is provided only on the title page (fol. 1r), which reads "Fingal’s Cave / Mendelssohn" and bears the stamp of the Philharmonic Society of London.

Copy CL includes rehearsal letters whose occurrences align with those in the London autograph (source AL).

CK. Badische Landesbibliothek, Karlsruhe, Mus. Ms. S. B. 8 Nr. 4.[74] Source CK is a set of nineteenth-century manuscript parts for Op. 26, bound within a collection of orchestral parts for a variety of concert overtures (including also Mendelssohn’s Opp. 21, 27, and 32). These parts are in the handwriting of two copyists, and are demonstrably derived from the first German edition of the score (1835; see discussion of ES under the print sources, below). They are of no significance regarding the Overture’s genesis or early performance history, but they do serve as evidence that the textual variants preserved in that first score edition continued to circulate in the later nineteenth century.

CX. Copyist’s parts used for the performances in London, 14 May and 1 June 1832 (lost). Since the textual content of CK indicates that those parts were generated from the printed score, there must have been another set of manuscript parts that were used for the 1832 London performances, and perhaps also for the other three performances given before the publication of the orchestral parts. These parts are now lost, but as noted above, their content appears to be transmitted through CL.

Print sources[75]

EPf1. First edition of Mendelssohn’s arrangement of the Overture for piano duet. This edition was published simultaneously by Breitkopf & Härtel (Leipzig) and Mori & Lavenu (London). As noted above, the release date was 15 October 1833. As revealed by the composer’s correspondence, the German edition was evidently prepared on the basis of English plates. The title page reads: OUVERTURE / aux Hébrides / (Fingals Höhle) / composée et dédiée / à Monsieur François Hauser / par / F. Mendelssohn-Bartholdy / arrangée / Pour le Pianoforte à quatre mains / par / L’ AUTEUR. / Propriété des Editeurs. / Leipzic, chez Breitkopf & Härtel et Londres, chez Mori & Lavenu. / Pr 1 Thlr / Enrégistre dans les Archives de l’ Union." There is no plate number on the title page, but the pages of the score bear the plate number 5483.

EP. First edition of the orchestral parts. Published simultaneously by Breitkopf & Härtel in Leipzig and Mori & Lavenu in London, these parts were released in late March 1834. Despite Mendelssohn’s statement that "[their] excellent production [left] nothing to be desired,"[76] the publisher did fail to observe at least one of his requests, for none of the parts contains all the rehearsal letters whose inclusion he had specifically requested in his letter of 29 November 1833.[77]

EPf2. Arrangement for Piano Solo by Friedrich Mockwitz, published by Breitkopf & Härtel in the fall of 1834. The title page reads: Ouvertüre / zur / Fingals-Höhle (Hebriden) / von Felix Mendelssohn-Bartholdy / für das Piano-forte arrangirt / von F. Mockwitz. The edition bears the plate number 5532.

As noted above, Mendelssohn had nothing to do with the production of this arrangement and disapproved of it generally; consequently, it is no relevance in the stemma of primary sources vested with the composer’s authority. Worth noting, however, is that it presumably was generated from the composer’s own arrangement for piano duet (APf).

ES. First edition of the full orchestral score.[78] Sent to Mendelssohn on 11 March 1835,[79] released in mid-March 1835,[80] and reissued ca. 1845, this edition is the last major print source of the Hebrides Overture issued during Mendelssohn’s lifetime. While Breitkopf & Härtel held the rights to the German editions of all three overtures in the triptych, the copyright for the English editions varied: the Midsummer Night’s Dream Overture was published in London by Cramer, Addison & Beale, while the rights to the Hebrides and Meeresstille were sold to Nicholas Mori.

As shown above, Mendelssohn conveyed to Breitkopf & Härtel his wish that the Hebrides share a single opus number with the Midsummer Night’s Dream and the Meeresstille und glückliche Fahrt Overtures, and that each work have its own plain title page as well as a collective title page (letter of 29 November 1833). These directions were imperfectly realized: the overtures were provided with a collective title page, but each did not receive its own individual title page. As in the instance of the orchestral parts (see above), ES failed to comply with certain specifications given in Mendelssohn’s letter of 29 November 1833; indeed, the score includes no rehearsal letters at all.

ER. Edition of orchestral score published in Felix Mendelssohn Bartholdy’s Werke: Kritisch durchgesehene Ausgabe von Julius Rietz (Leipzig: Breitkopf & Härtel, 1874-77), series 2. As is well known, the Rietz editions of Mendelssohn’s works, despite laudable intentions, possess serious problems of philological credibility. The set is sometimes referred to informally as a Gesamtausgabe, yet its presentation of Mendelssohn’s oeuvre is far from complete. Moreover, the editions themselves are in many ways corrupt. Most often, they are tacit conflations of the texts as offered in the autographs known to Rietz, editions released prior to the 1870s (including some with which the composer was not involved), and the editors’ judgment calls. This is true also of Rietz’s edition of the Hebrides Overture: it is a tacit conflation of the first score edition, the London autograph (source AL), and other unaccountable editorial decisions.

The above description of the extant sources clearly emphasizes the Hebrides Overture’s convoluted gestation – but a comparison of the extant sources with those whose existence is documented by chronographic and circumstantial evidence also confirms that still more sources are missing or have been lost.[81] There must have been proofs at least for the first edition of the arrangement for piano duet and the first orchestral score, but these have not survived.[82] The score Mendelssohn presented to the Philharmonic Society of London on 6 June 1832 is also lost, as are the parts that were used for the 1832 London performances. Moreover, given the extensive revisions the Overture underwent between those performances and the Berlin performances of early 1833, a second set of parts probably was also generated for the latter. These parts, too, are lost – indeed, they evidently were lost already by fall 1833, when Mendelssohn and his Berlin family were unable to locate them for forwarding to Breitkopf & Härtel.[83] Also not clearly accounted for is the score Mendelssohn sent to the Leipzig firm on 29 November 1833, which was to serve as the basis for the first edition of the orchestral parts.

Examination of the surviving sources makes it possible to fill some of these lacunae. To begin with, as Peter Ward Jones has shown, the text of the work as it existed in the London 1832 performances is most likely transmitted through the copyist’s manuscript (CL) prepared by William Goodwin, who worked as a copyist for the Philharmonic Society from at least 1829, and who later became its assistant librarian.[84] Source CL diverges substantively from both AR and AL but transmits significant features of both (see discussion below).

The latter observation is significant in several regards. First, it suggests that at least some of the revisions transmitted in source AL date from after that score’s nominal date of completion (20 June 1832). Moreover, if AL served as the basis of the proofs that were generated in fall 1834 and returned, with "still more corrections," by Mendelssohn on 15 November 1834, then AL was almost certainly returned to Mendelssohn along with the proofs based on it; the composer then probably would have entered his subsequent changes to the proofs into his own manuscript as well. In other words, some revisions in source AL may well date from as late as mid-November 1834 – nearly two and a half years after that autograph’s nominal date of completion (see discussion below).

.

III. The Revision History

The above remarks describe a complex of sources whose interrelationships are obviously close – but even these cursory descriptions offer glimpses into the myriad chronological contradictions and ambivalences that complicate any effort to gain a clear view of the work’s genesis or formulate chronological guidelines for textual authority. For example, according to the dates given at the end of the autographs the composer’s piano-duet arrangement (APf) pre-dates the second autograph full score (AL) by one day – but as will be shown presently, APf tacitly reflects the removal of two sizeable groups of measures that are deleted with cross-hatching in AL, so that the total number of measures in the piano arrangement concurs with the post correcturam reading of the orchestral score supposedly completed a day later.

Such paradoxical relationships among the sources certainly complicate the task of sorting out the various stages in the Overture’s gestation, but the obstacles they present are not insurmountable. On the contrary, the rough chronology of the sources offers a starting point for a closer examination of the revisions that articulate the work’s successive versions. Since manuscript copies CO and CL (and perhaps also CD) have a clearly defined provenance and close proximity to the composer, we may provisionally treat those copies as essentially faithful reproductions of the autograph sources on which they are based, as those autographs existed at that time. Moreover, if we assume that Mendelssohn would have been less likely to vacillate on large-scale revisions than on incidental or local ones, we may securely correlate the available information concerning the overall disposition of the sources with more specific information concerning large-scale revisions. This correlation in turn offers suggestions as to the chronology of these large-scale textual emendations reflected in the various manuscripts. Finally, with the chronology of the large-scale revisions generally established, we may review smaller-scale alterations. The result is a remarkably detailed chronology of the particular contents of the text as it existed at various points in the Overture’s gestation.

Large-Scale Revisions

The large-scale issues that occasioned Mendelssohn’s protracted reworking of the Hebrides Overture are revealed in eight reworkings of passages transmitted by the two autograph full scores (AR and AL). The source-to-source textual correlations among these large-scale revisions, collectively described as the deletion of one or more measures or the wholesale replacement of material from the earlier autograph with different material in the later one, are summarized in Table 2. A more detailed explanation follows:

No. 1. End of the principal subject and transition to the second subject, corresponding to mm. 30-47 in the final score. In AR, this passage comprises thirteen measures, extending from the third bar of p. 5 through the eighth (final) bar of p. 6; as noted above, Moscheles later on framed this passage in large "X"s. The revised reading, largely corresponding to the final version, is transmitted on 4r (m. 3) to 5r (m. 4) of AL. Manuscript copies CO and CD concur with AR in their reading of this passage, while the autograph piano-duet arrangement (APf) and copy CL concur with AL.

As shown in Example 2,[85] the original version of this passage clearly conformed to the expectations of a transitional passage in sonata-allegro form: an authentic cadence in the tonic B minor in m. 30 was followed by an immediate move toward the minor dominant, coinciding with metrically alternating statements of the Kopfmotiv – an increase in non-cadential sequential motion and rhythmic fragmentation characteristic of this passage in conventional versions of the form – followed by the arrival on the dominant of the second subject’s key, D major. As R. Larry Todd has observed, this transition is characterized not least of all by its wealth of imitative textures.[86] In its final version, the transition is less conspicuous in its artfulness: the harmonic strategy is simpler, the texture sparser, and the sense of a telos significantly reduced. Since the learned devices of AR are inconsistent with the rustic atmosphere of the Overture’s poetic subject as Mendelssohn conceived it at that point,[87] it seems clear that this first large-scale revision arose at least as much because of poetic concerns as musical ones.

No. 2. Closing section of the exposition, corresponding to mm. 70-96 in the published score. In AR, this passage comprises 38 bars, extending from the third measure of p. 10 through the penultimate (eighth) measure of p. 14. As with the first large-scale revision, Moscheles inscribed a large "X" in pencil at the beginning and ending of this passage (see Fig. 4, above). As shown in Example 3a, this passage clearly is the one Mendelssohn had in mind when he wrote to Fanny on 21 January 1832 that "[t]he middle part in D major marked forte is rather ridiculous, and the entire, so-called working-out tastes more of counterpoint than of train oil, gulls, and salted cod – it should be just the other way around."[88]

As noted above, the composer’s correspondence documents that this passage had been replaced by 5 May 1832, and the replacement passage is indeed transmitted in autographs APf and AL. However, those two sources do not concur exactly in their readings of the replacement music; for as shown in Examples 3b and 3c, AL includes six measures between what would become mm. 84 and 85 of the familiar version. In AL (Ex. 3b) these six bars are deleted via cross-hatching; in APf and CL, however, they are missing altogether (Ex. 3c). This clearly indicates that the deletion of the six measures in AL occurred prior to the writing-out of APf, even though the date on its last page suggests that APf was completed first.

Copies CO and CD concur with AR in their presentation of this passage, while copy CL concurs with autograph APf and, on the whole, the post correcturam reading of this passage in AL (see discussion below).

No. 3. New theme in the development section, corresponding to mm. 112-23 of the final version. As shown in Figure 7, AR originally contained four bars between what became mm. 114 and 115 of the published score; these four bars comprised the last two on p. 16 and the first two on p. 17. In the length of the passage as a whole, all other manuscripts concur with AL and differ from AR; that is, the measures in question are missing from all other manuscript sources. This clearly indicates that the deletion was accomplished prior to 25 February 1831.

No. 4. Development section, corresponding to mm. 131-47 of the final version. The autographs’ readings of this section of the development, beginning just after the return of the second subject in D major, differ primarily in length (27 measures in AR as opposed to 17 in AL) and in developmental intensity. As in the first of the large-scale revisions discussed above, the developmental technique of the original version was more expansive and texturally elaborate, while that of the revised reading is sparer and less reliant on conventional developmental techniques involving imitation and sequence (see Example 4a-b). Two of the manuscript copies, CO and CD, concur with AR, while the third one, CL, concurs with the ante correcturam reading of AL (see discussion below).

No. 5. Retransition, corresponding to mm. 169-74 of the familiar version. In AR, this passage extends from the third measure of p. 25 through the sixth measure of p. 26; the remaining five bars of the retransition correspond to the familiar version. Copies CO and CD transmit this passage as given in AR (Ex. 5a), but the relationship between AL, CL, and APf is more complicated. As shown in Example 5b, AL originally contained two measures between what became mm. 174 and 175, and the deletion of these measures is tacitly reflected in both CL and APf (see discussion below).

No. 6. Extension of the principal subject in the recapitulation, corresponding to mm. 188-93 of the final version. In AR this passage comprised the last three bars of p. 28 and all of p. 29 – a total of ten measures that would be halved in the course of the revision (cf. Ex. 6 and mm. 186-93 of the familiar version). Here as with revisions 3 and 4, copies CO and CD concur with AR. Copy CL and autograph APf, by contrast, concur with AL (whose text corresponds to the familiar version). The lateness of this important revision is noteworthy because the material given in AR belongs to some of Mendelssohn’s earliest sketches for the work, evidently dating from mid-October 1830.[89] Mendelssohn at a very late date aborted material that had been preserved intact through numerous revisions dating from early on in the Overture’s genesis.

No. 7. Coda, between mm. 221 and 222 of the published version. As shown in Figure 8, page 33 of AR transmits a cross-hatched measure between mm. 221 and 222. Manuscript copies CO and CD present the measure intact (i.e., those sources give the ante correcturam version of this passage), while the measure is absent in CL, APf, and AL.

No. 8. Coda, between mm. 231 and 244 of the published version. As shown in Example 7a, mm. 232-233 of the published version are a the result of a substantial revision accomplished already in source AR. Page 34 of AR gives the essential content of mm. 226-31 as it is known from the familiar version, but between mm. 231 and 244 of the published version AR gives a total of twenty-five original measures, of which the last four on p. 37 were deleted. Copies CO and CD retain the ante correcturam reading of this passage. In the corresponding passage in AL Mendelssohn retained individual gestures from the post correcturam reading, with modifications, tacitly leaving behind other measures (most notably, the rather vacuous material that comprises the bulk of p. [26]). The result is that the published version accomplishes in thirteen measures what took twenty-five measures to accomplish in AR (see Ex. 7b).

The above observations permit some provisional deductions about the chronological relationships among the sources. To begin with, of the eight large-scale revisions reflected in the extant manuscripts, only one (No. 3) had been accomplished by 25 February 1831 (the date on which the composer described the "italienische Notenschreibesprache," clearly referring to CO, to Fanny). That date thus serves as a terminus post quem for the remaining seven large-scale changes reflected in AR. This observation sheds new light on the date of 16 December 1830 given at the end of the score, and on Rome as locus of the music present in AR. While 16 December 1830 represents the initial date of completion for AR, that manuscript evidently served as Mendelssohn’s working manuscript for some time afterwards, probably as late as April 1832. In fact, most of the large-scale revisions present in AR (nos. 1, 2, 4, 5, 6, 7, and 8) clearly date from between 25 February 1831 and early 1832 (after which point AL supplanted AR as Mendelssohn’s Arbeitspartitur).

The dates of 19 and 20 June 1832 at the end of APf and AL must likewise be taken cum grano salis, for APf tacitly incorporates two large-scale revisions (Nos. 2 and 5) accomplished via cross-hatching in AL even though it was supposedly completed a day before AL. Moreover, since CL was evidently scored up from the parts used for the London performances of 14 May and 1 June 1832, we may assume that the cross-hatching of those measures in AL occurred prior to the writing-out of lost autograph AX (the source of the parts from which CL was reconstructed). Since as a practical matter AL must have existed as a manuscript in order to serve as the basis for AX, we may conjecture that the notation and crossing-out of passages Nos. 2 and 5 occurred between 30 April 1832 (when Mendelssohn played the work for Moscheles) and 6 May (when he recorded "Ouv. fertig" in his diary and gave AR to Moscheles).

These observations lead to a tantalizing conjecture – namely, that in early May 1832 Mendelssohn must have been working at turns on three separate autographs for the Overture, perhaps using AL as the central revision score while also working on APf and the lost autograph (AX) from which the parts used in the premiere were generated. Two identifying features of the latter may be securely postulated. First, it would likely transmit the post correcturam reading of large-scale revisions nos. 2, 5, and 8, since AL graphically reveals Mendelssohn working on these passages but they are tacitly carried over into both APf and CL. Second, since Mendelssohn’s diary clearly records that one autograph score was completed on Sunday, 6 May, and no surviving autograph bears that date, the score in question may well have been AX. In this scenario, Mendelssohn would have given that score to the Philharmonic Society exactly one month later, on 6 June.

This raises the question, however, of when Mendelssohn began writing the London autograph score (AL) and the piano-duet arrangement (APf) – for clearly all the large-scale revisions in AL had been accomplished by the time APf was written out. Here, however, the paper of AL and APf provides useful information – for as noted above, both autographs are written on paper of obviously English origin. Since Mendelssohn arrived in London for the first time since 1829 on 23 April 1832 and on 30 April played an early version for Moscheles, AL almost certainly was written between 30 April and 5 May (the date on which the composer reported to his family that he had to write out the score again). Given the clean state of copy in APf, we may surmise that it dates from relatively late in that date-range.

Finally, circumstantial evidence and the presence of rehearsal letters in AL point to that autograph as the score Mendelssohn sent to Breitkopf & Härtel on 29 November 1833 for use as the basis of the printed orchestral parts. The composer clearly specified that this score contained carefully placed rehearsal letters, and his frenetic schedule between June 1832 and November 1833 makes it unlikely that he would have written out the score yet again in the interim (nor do his letters and diaries suggest such an undertaking). Such letters are present in three of the surviving manuscript scores: CD, CL, and AL. The score in question could not have been CD because its content replicates that of AR and CO, both of which were supplanted by AL already in May 1832. Nor do CL and AX appear to be viable candidates, since the former was later scored up from parts and remained in the possession of the Philharmonic Society, and the latter was given to the Society as a gift. AR thus emerges as the only reasonable alternative. This circumstance is significant, for as noted above Mendelssohn conducted the Overture at two performances in January and February 1833 – and for these occasions it seems likely that he introduced further changes that affected the text of the work, even though they were presumably of relatively minor significance.

By correlating these complicated interactions between autographs and copies with the known chronology of the Overture, we may securely assign specific ranges of dates to specific sets of revisions transmitted by the various manuscript sources for the Hebrides.

• Although the Rome autograph (AR) remained in use as Mendelssohn’s Arbeitspartitur for as much as sixteen months beyond its nominal date of completion, the text of that earliest complete version as it existed between December 1830 and late February 1831 is reliably transmitted by copy CO.

• Of the large-scale revisions discussed above, only No. 3 was present in this early reading.

• Alterations present in AR but not accounted for by CO may be generally taken to transmit the text of the work as it existed between February 1831 and ca. 30 April 1832 (the version that Mendelssohn reportedly played in Rome for Berlioz, in Paris for Ferdinand Hiller, and in London for Moscheles). Among these alterations are large-scale revisions nos. 7 and 8, as well as no. 3 (which had been accomplished by 25 February 1831).

• Until the lost autograph AX, presented to the Philharmonic Society on 6 June 1832, is recovered, the text of the work as it was offered in London on 14 May and 1 June 1832 is best transmitted by the copy of the score held in the British Library (source CL).

• Of the large-scale revisions discussed above, Nos. 1, 2, and 6, as well as the post correcturam reading of No. 8, were accomplished after the production of CO and CD but before Mendelssohn undertook to write out AL; i.e., sometime between 25 February 1831 and late April 1832 (since Mendelssohn evidently began work on AL in early May).

• Nos. 4 and 5 were accomplished after Mendelssohn had undertaken to write out AL but before the completion of AX, CX, and CL. These two large-scale revisions may thus be assigned a date of May 1832.

• The London copy (CL) makes clear that the rehearsal letters found in AL and the first edition of the parts were in place by the summer of 1832.

• Since No. 7 is tacitly accounted for in both CL and APf, but was graphically accomplished in AL, that revision must pre-date the production of those two manuscripts.

• Alterations represented in the latest autograph score (AL) but not accounted for in CL may be presumed to have been entered into AL sometime between the London performances (May-June 1832) and the delivery of AL to Breitkopf & Härtel for engraving of the orchestral parts (29 November 1833).

• If the manuscript orchestral parts used for the Berlin performances in early 1833 ever turn up, these will in turn permit differentiation of readings in AL that existed before and after those performances. Since those parts were lost already later in 1833, however, this seems a remote possibility.

• Finally, variants between the orchestral parts and the first edition of the orchestral score may generally be assigned to the period of mid-August to mid-November 1834.

In summary, the large-scale revisions document that the pre-publication textual genesis of the Overture falls into four discrete stages: (1) from December 1830 to 25 February 1831; (2) from 25 February 1831 to 30 April 1832; (3) from 5 May to ca. 6 June 1832; and (4) from ca. 6 June 1832 to ca. November 1833. Moreover, variants also occur between the orchestral parts (produced between November 1833 and early April 1834) and the first edition of the orchestral score (proofread between August and November 1834 and published in late March 1835). From a philological and textual perspective, therefore, it is awkward to refer to the work in only three discrete versions ("Rome," "London," and "final" or "familiar"). If such reference is made, it should reflect CO, CL, and ES rather than the Mischfassungen transmitted by AR and (for example) AL or ER; preferable would be to refer to "Rome," "Rome-Paris," "London," "London-Berlin," and "Berlin-Leipzig" versions.

Even in this instance, however, the editor and performer must remain cognizant that each of these five chronologically discrete versions of the work articulates only a stage of revision manifested in large-scale revisions. The manuscripts also transmit a number of "local" revisions (changes that affected neither the length nor the essential content of the Overture), most of which are linked to these large-scale revisions only incidentally. To identify all these local revisions would exceed the scope of this study. The following remarks examine some such revisions that are significant for the music and our understanding of its genesis.

Local Revisions

Four groups of local revisions warrant special attention in these pages: a set of revisions in AR that may now be assigned to specific periods in the Overture’s genesis; an instance centering around practical issues of instrumentation; a set of revisions in AL that may now be chronologically pinpointed; and, finally, an instance of local revisions that also relate to broader motivic and thematic issues.

The first of these groups hinges on the postulation of ca. 25 February 1831 as the terminus ante quem for revisions incorporated into the Oxford copy (CO) from the Rome autograph (AR). As noted above, No. 3 of the large-scale revisions discussed above may be securely dated as having been accomplished between December 1830 and late February 1831; conversely, the remaining seven large-scale revisions in AR were accomplished only after late February 1831. By contrast, a comparison of AR and CO reveals that all but a very few of the local revisions evident in AR had been accomplished by the time CO was copied, for these local revisions are consistently given in their corrected form in CO. Moreover, those few local revisions present in AR but not accounted for in CO occur only in the outermost pages of the manuscript (transmitting music that corresponds to mm. 1-27 and 265-68 of the published version).

Two representative passages will suffice to illustrate this situation. In mm. 19-20 (AR p. 3), three local changes were entered: the entry of the flutes on the pickup to measure 20 was changed from to ; the originally notated pitch D was changed to D in the second violins and violas; and in the Violin 2 staff the fourth beat was revised to comprise a single eighth-note followed by an eighth rest (see Fig. 9). All these changes are tacitly accounted for in CO, indicating that they were accomplished before 25 February 1831. In mm. 1-5 (AR p. 1), however, the second bassoon originally sounded on the first three beats of each measure (see Fig. 3 above). These entries are preserved intact in CO even though they were at some point deleted in AR. Assuming that CO is a reasonably accurate representation of the Overture’s text as it existed in late February 1831, this latter revision must indicate that the second bassoon part in mm. 1-5 was deleted only after the production of CO. Since the version of the work that the composer played for Moscheles on 30 April 1832 was almost certainly the one represented in AR, the deletion of the second bassoon part may be assigned to some point in those intervening months.

Again, these chronological assignments affirm that the date of "Rome, 16 December 1830," given at the end of AR says but little at all about the revisions contained in the autograph. For certain we can say only that they were in place before 5 May 1832, since there is no sign of them in CL and AL. What is more, the relative paucity of large-scale revisions reflected in CO, along with that manuscript’s consistent inclusion of local changes present in AR, suggests that until late February 1831 Mendelssohn was generally satisfied with the large-scale structure of the Overture and up to that point introduced mostly minor revisions. In the case of the Hebrides, the Revisionsteufel (as he termed it) that led to extensive revisions and deletions set in only sometime later on.

The manuscripts also reveal another local change of a more practical import. As most of the examples and figures given so far reveal, Mendelssohn originally conceived that for the first half of the work the clarinets would be pitched in C; only in the reprise did he specify the change to clarinets in A (see Fig. 10). This change probably was occasioned by the timbral warmth called for by the return of the second subject in B major, entrusted to the soli clarinets, in what would become m. 202. This usage of the two instruments was originally retained in AL. In that manuscript, the clarinets are notated in C through m. 187; in m. 188 (fol. 13v) Mendelssohn specified a change to clarinets in A, twelve bars before the solo entry. Immediately after the solo, on beat 3 of m. 217, AR then specifies "Clarinetti in C," preparing for the instruments’ reentry on the last eighth note of m. 221 – a remarkably (and perhaps impracticably) quick change. A similar situation obtains on the final leaf of the score (fol. 18r-v): in m. 258, after just under two measures of rest, the C clarinets revert to A clarinets – and in this instance the autograph reveals Mendelssohn’s compositional decision-making even more clearly: as shown in Figure 11, the clarinet staff on p. 18r was initially provided with the key signature for instruments in C, but Mendelssohn then changed the signature of two sharps to one of one flat (i.e., for instruments in A), inserted the directive "in A," and notated the entire page (and the four bars on fol. 18v) at that transposition. At some point afterwards, however, he must have realized that this change was even more difficult than the previous one. It was presumably at this point that he added the further directive "gilt" ("stet") next to the clarinet staff – presumably as a warning to a copyist or engraver.

The difficulty of these improbably quick changes of instruments sheds light on Mendelssohn’s practicality as an orchestrator, and in this instance, too, source-to-source textual variants lend insight into the chronology of the sources. For while the main Notentext of AL offers no way around these orchestrational difficulties, the first page of the manuscript reveals that at some point afterwards he thought better of the situation: above the opening measures of the clarinet staff is the inscription "müssen durchgängig in A stehn" ("must remain in A throughout"; see Fig. 5, above). Moreover, this directive is tacitly accounted for in CL: the clarinets are notated in A throughout. Returning to the chronology sketched above, this variance would suggest that Mendelssohn realized the difficulty of his original specifications sometime between the initial completion of AL (6 May 1832) and, at the latest, the premiere on 14 May. Most probably, the difficulties came to light during the rehearsal (12 May) and the composer decided then to use clarinets in A throughout. Although it is conceivable that this on-the-spot revision could have been accommodated by the players and specified into their parts (CX), so that the copyist of CL (Goodwin) tacitly adjusted the entries in the score when producing it later on, such a conjecture must await the reemergence of those parts (or of lost autograph AX from which they were extracted).

At the same time, a comparison of AL, CL, and EP permits the assignment of numerous revisions to the period between the nominal completion of AL (20 June 1832) and the point at which the composer sent the corrected proofs for the parts to Breitkopf & Härtel (29 November 1833). For example, the string parts on fol. 18r indicate that Mendelssohn continued to use AL, and to revise it substantively, beyond the production of the parts from which CO was generated (CX). In AR (as in CO and CD) the strings essentially replicated the melodic and rhythmic distribution of the winds (Ex. 8a). This colla parte relationship forms the ante correcturam reading of the passage in AL, and it evidently was preserved in lost autograph AX and the parts used for the 1832 London performances, for CL gives the same reading. At some point thereafter, however, Mendelssohn revised the passage to read as it does in the familiar score (Ex. 8b). Potentially, the window of opportunity for the revision could extend to 29 November 1833, since the corrected reading is present in EP (the first edition of the parts).

A more complicated instance obtains with regard to large-scale revision No. 2 (the passage that Mendelssohn referred to as "matte Lärmmacherey"). As noted above, sources AR, CO, and CD transmit the early version of this passage, while on the whole APf and CL concur with the post correcturam reading of AL – i.e., the deletion of six measures on fols. 7v-8r is tacitly accounted for in APf and CL. At the same time, however, AL gives two distinct sets of local revisions in this passage: several that are accounted for in APf and CL and several that are not. To the former group belong revisions in mm. 74 (clarinet 1), 75 (horns and violas), 76 (violins and violas), 77 (oboes and clarinets), 78 (horn 2), 79 (violin 1), 80 (clarinet 1, cello, and bass), 82 (horns, cello, and bass), 84 (horns), 85-86 (all woodwinds), and 85 (timpani and strings). The revisions that are not accounted for in CL and APf occur in mm. 72-73 (cello), 80 (violins and violas), 84 (violin 2 and violas), and 87-88 (violins and violas). Clearly, the first of these groups of revisions pre-dates the production of CX (the copyist’s parts from which CL was generated), and may thus be dated between 5 May and 6 June 1832. The second group, by contrast, must date from between 6 June 1832 and mid-March 1834.

Finally, one local change exerted a pronounced influence not only on the Overture’s affective content, but also on its motivic fabric. As is well known, one cardinal feature of the work is its pronounced motivic unity – or, to put it differently, its demonstration of Mendelssohn’s imagination in weaving a complex and expansive musical tapestry out of surprisingly few musical ideas. The early versions (up through APf and EPf2) all featured one motivic gesture whose subsequent replacement enhanced this motivic unity, for from AR through APf and even EPf, the insistent 

repeated-note figure in mm.169-70 was not the familiar
motive, but the much tamer dactylic figure
(see Exx. 5a-b, above).
When Peter Ward Jones discussed this revision in 1997 source AL had yet to turn up at auction.[90]  Now, however, it is possible to examine the passage in question – and indeed, the autograph vividly reveals Mendelssohn’s decision-making process (see Fig. 12): in the winds, brass, timpani, and upper strings alike, irregular stem-lengths and spacing reveal that the familiar rhythm was written to accommodate the already-written note-heads and stems of the dactylic rhythm from the earlier version. The dactylic rhythm is present in the ante correcturam reading of AL as well as CL, APf, and even EPf1 and EPf2, but the revision is accomplished in EP. The possibility thus exists that Mendelssohn’s entry of this significant alteration occurred sometime in the fall of 1833.
The above observations not only permit a detailed reconstruction of the textual genesis of the Hebrides Overture, but also encourage the scholar to group the textual variants into five discrete classes, each representing a textually stabilized moment in the work’s genesis and many potentially viable as correspondingly classified ossia readings.[91]  The first two classes of variants hinge on the identification of 25 February 1831 as the date at which CO assumed the status of a reliable copy of AR: to the first of these classes belong the "Rome" variants (present in both CO and AR); to the second, the "Rome-Paris" variants (those present in AR but not CO). The next two classes hinge on the textual relationships between CL and AL: variants present in both CL and AL comprise the "London" variants, while those present in AL but not in CL may be described as "London-Berlin" variants. Of greater chronological authority are the "Berlin-Leipzig" variants (those that appear for the first time in EP), since these reflect changes that Mendelssohn presumably entered in response to the two performances in Berlin (10 January and 14 February 1833) and a further performance in Leipzig (13 February 1834). Finally, there are the "Leipzig" variants – those that appear for the first time in ES. This latter group, which reflects the composer’s latest ideas on the ideal state of his work, includes the variant of mm. 7-8 given in Example 1b, above. Indeed, while that particular member of the latest variants transmitted by contemporary sources ties the Overture’s introduction and principal subject to the "new" theme first presented in the development,[92]  it also serves a purpose more historical and biographical in nature. For as the composer’s letter of 7-11 August 1829 reveals, the integration of the consequent phrase into the introduction was a part of Mendelssohn’s earliest conception of the work (see Fig. 1, above). Far from being a last-minute vacillation, this revision enabled the composer finally to realize his original conception of the work – in a form, it must be admitted, significantly more successful than had earlier been the case. Unfortunately, however, most of these variants are not represented in most editions currently available.
. IV. Conclusions
What lessons do the protracted genesis and convoluted source-situation of the Hebrides Overture offer to those who venture into their entanglements? The question is well worth asking, for while this composition is exceptional in some ways, it is typical in several others – and these representative features are highly instructive. To begin with, the Hebrides is a case-study in the working methods and compositional processes of a composer who continuously revised his compositions up through the latest stages in publication process (and sometimes beyond). Mendelssohn’s clear script, which offers but few of the extremely illegible notations characteristic of (for example) Beethoven, often masks a vast and tangled jungle of complete, partial, and palimpsest-ridden revisions. The sources and the works’ geneses reveal that he struggled mightily in composing; that he was more relentlessly critical of his own works than others were; and that his self-critical faculties accounted for his renunciation or suppression of any number of works that many composers would have been proud to claim as their own. These findings – the results of a genre of scholarly discourse that in Mendelssohn’s case was opened up only relatively recently[93]  – are significant not least of all because they refute the conventional late-nineteenth-century image of Mendelssohn as a composer possessed of a certain glib facility and a proclivity for superficial formalism. In the case of the Hebrides Overture, the manuscript sources reveal him grappling with major musical issues long after the work had won the admiration of Berlioz, Hiller, and Moscheles. In particular, the extremely late revision of mm. 7-8 demonstrates that his efforts to improve his music continued even beyond the point at which the work had achieved resounding success from the critics and the public in publication as well as performance. Perhaps more important, especially for readers of this journal, is the insight that the sources for the Hebrides offer concerning a point that is all too often missed: the fact that the literal chronographic evidence often is also quite misleading. In this instance, the dates the composer appended to the two autograph scores, like the entries he provided in his diary, ascribe to those manuscripts dates that ultimately are only of limited use in understanding the work’s chronology. Both AR and AL include substantial amounts of material that came into existence only after their respective Schlussdaten, which concur with Mendelssohn’s diary entries. Without a close philological inspection of the autographs in conjunction with contemporary copies and early print editions, those self-evidently chronographic indicia would offer misleading perceptions of the Overture’s genesis. What is more, the philological and textual challenges posed by the Hebrides are emblematic of the difficulties that have long thwarted efforts to produce source-critical editions of Mendelssohn’s works. Scholars’ increasing grasp of the extant sources, the ways in which these relate to each other and to other sources that are now lost or missing, and the relationship of both to the daunting body of contradictory and vague chronographic indicia not only reaffirms the incompleteness and editorial inadequacy of the compositions as they are presented in the ostensibly "critically reviewed" edition of Mendelssohn’s Werke published by Breitkopf & Härtel in 1874-77. It also underscores the irony of the fact that those editions continue to constitute the central sources by which Mendelssohn’s music is performed and studied.[94]  Awareness of the problems has generated at least three major scholarly initiatives aimed at providing reliable editions of Mendelssohn’s music: the Leipziger Ausgabe der Werke von Felix Mendelssohn Bartholdy, now moving ahead rapidly under the general editorship of Christian Martin Schmidt; a separate series of the sacred works published by Carus-Verlag (Stuttgart) between 1977 and 1997; and, most recently, a series of authoritative editions of the major concert overtures, edited by Christopher Hogwood and published by Bärenreiter (Kassel). The latter group is of particularly felicitous relevance to this article, for in 2004 Professor Hogwood’s edition of the Hebrides was released – an edition that not only is the first to identify the variants in a critical report, but also presents many of them as ossia readings.[95]   Most important of all, however, are two further remarkable opportunities exemplified by the textual and philological problems of the Hebrides. First, for the editor who wishes to present the Overture in any one of its chronologically stabilized textual states prior to the Fassung letzter Hand represented by ES, the most appropriate copy-texts are not the autographs – for as shown above, all the autograph scores are Mischfassungen. Rather, only the contemporary copies (CO, CD, or CL) may be regarded as representative of the text in chronologically stabilized forms. This corpus of reasonably pure copy-texts for the early texts of the orchestral version is supplemented by the first edition of the orchestral parts (EP), which transmits the text of the work as it existed in November 1833, and by ES, which transmits the composer’s latest thoughts on the Overture’s revision. The evident circulation of ES in the later nineteenth century, along with the proliferation of other scores and arrangements derived from that edition, certifies that for most of that century an edition that represented the composer’s final authorized version of the Hebrides did serve as the starting-point for performers and commentators. When the Rietz editions appeared, however, they tacitly combined some readings from ES with others from AL and still others whose provenance remains unclear. As scholars and performers have accepted the putative authority of the Rietz editions, the text of the Overture as finally authorized by Mendelssohn went out of circulation; coterminously, the textually purer editions previously in circulation were subverted to the unreliable text presented in the Rietz edition. In other words, since the late nineteenth century most performers and scholars have known the Hebrides Overture only in a textually corrupt state. Finally, if the Hebrides Overture is edited according to editorial ideology of Fassung letzter Hand, the copy-text should be the first German edition of the score, and the autographs should be used only to verify corrections to obviously erroneous readings transmitted therein. In this instance, however, that methodology would result in a substantial alteration of the work’s text as it was known through 1834, and it has been known since the appearance of the Rietz editions in the 1870s. For in the Fassung letzter Hand of this work the memorable gesture in the familiar version of mm. 7-8 (Ex. 1a) would have to be stricken and replaced by the substantially different reading Mendelssohn provided de ultissima hora (Ex. 1b). Such difficult philological imperatives are common in Mendelssohn’s oeuvre. But the challenges also have their rewards, for a steadily increasing awareness of the problems with the editions by which Mendelssohn’s music has traditionally been known continues to lend decisive momentum to the current resurgence of scholarly and musical interest in the composer.
________________________
[1] Julius Rietz, ed., Felix Mendelssohn Bartholdy’s Werke: Kritisch durchgesehene Ausgabe (Leipzig: Breitkopf & Härtel, 1874-77). [2] See R. Larry Todd, "The Instrumental Music of Felix Mendelssohn-Bartholdy: Selected Studies Based on Primary Sources" (Ph.D. diss., Yale University, 1979); idem, "Of Sea Gulls and Counterpoint: The Early Versions of Mendelssohn’s Hebrides Overture," Nineteenth-Century Music 2 (1979): 197-213; and, most recently, Mendelssohn: "The Hebrides" and Other Concert Overtures (Cambridge: Cambridge University Press, 1992), esp. 26-37. See also Andreas Eichhorn, Felix Mendelssohn-Bartholdy: Die Hebriden, Ouvertüre für Orchester op. 26, Meisterwerke der Musik, Hft. 66 (Munich: Wilhelm Fink, 1998), esp. 7-17. [3] See Todd, "The Instrumental Music," and idem, "Of Sea Gulls and Counterpoint." [4]  For a reproduction of this drawing, see Todd, "Of Sea Gulls and Counterpoint," p. 205, or Eichhorn, Die Hebriden, Pl. II. [5] For a transcription of this sketch, see Todd, Mendelssohn: "The Hebrides," 28-29. [6] See Todd, "Of Sea Gulls and Counterpoint," esp. 204-08. [7] See Paul and Carl Mendelssohn-Bartholdy, eds., Felix Mendelssohn Bartholdy: Briefe aus den Jahren aus den Jahren 1830-1847 (Leipzig: Hermann Mendelssohn, 1864), 1: 69-71 (original in the Bodleian Library, Oxford, MS M. Deneke Mendelssohn d.13, fol. 25). [8] Paul and Carl Mendelssohn-Bartholdy, Briefe 1830-1847, 1: 81-83. [9] Although Todd surmises that this letter may have referred to an actual completion of the work on 11 December and surmises that this early version may be documented in the manuscript copy that today survives in the Bodleian Library (described as source CO later in these pages), the evidence suggests that the Bodleian copy post-dates the autograph completed on 16 December. [10] Berlioz’s recollections of his early acquaintance with Mendelssohn are vividly summarized in his Mémoirs, written and published more than thirty years later; for excerpts from these recollections, see Roger Nichols, Mendelssohn Remembered (London: Faber & Faber, 1997), 171-75. Understandably (given their chronological distance from the events in question), Berlioz’s remarks do not always square with independently verifiable facts and Mendelssohn’s own private recollections. The artistic and personal relationship between the two composers has yet to be explored in detail in print. The most reliable account to date is found in Professor Todd’s magisterial new biography of Mendelssohn (see R. Larry Todd, Mendelssohn: A Life in Music [Oxford: Oxford University Press, 2003]). Among other things, Todd points out (in contradiction to Berlioz’s statement) that the two composers saw each other again in June 1831 (p. 237-39). [11] See Peter Sutermeister, ed., Felix Mendelssohn Bartholdy: Briefe einer Reise durch Deutschland, Italien und die Schweiz, und Lebensbild . . . mit Aquarellen und Zeichnungen aus Mendelssohns Reiseskizzenbüchern (Zurich: M. Niehans, 1958; rpt. Tübingen: Heliopolis, 1979), 112-13. [12] Unidentified; probably "Emil B," who figures in one of Mendelssohn’s diaries from the Roman sojourn (MS M. Deneke Mendelssohn g.2). See Pietro Zappalà, "Dalla Spree al Tevere: Il Diario del viaggio di Felix Mendelssohn Bartholdy verso l’Italia (1830-1831): Edizione e commento," in Album amicorum Albert Dunning: In occasione del suo LXV compleanno, ed. Giacomo Fornari (Turnhout: Brepols, 2002).
[13]  See Sutermeister, Briefe, 112-13:  "Liebe Fanny! Dies ist die Hebridenouvertüre. bedeutet p, heißt sfz, heißt f. in der italiänischen Notenschreibesprache, sapienti sat. Du wirst am Ende des sogenannten ersten Theils, wo es in d dur schließt eine schlechte Stelle finden; ich habe sie ändern wollen, ehe ich es schickte, aber Zeit fehlte. Denke es Dir also anders. Die matte Lärmmacherey von da an u. die folgende Stelle, die aus meiner vortrefflichen Reformationssinfonie sichtlich abgeschrieben ist, u. mit der ich mir selbst eine Schmeicheley sage, sollen anders werden, sobald ich abkomme. Einstweilen nimm es so an. Ich schrieb es hin weil ich Eile hatte u. verschob die Anordnung, weil andere Arbeiten drängten. Emil reist morgen u. nimmt alles mit. Möge es Dir gefallen..."

[14] Letter to Fanny Hensel dated 21 Jan. 1832, published in Briefe 1830-47 (1864) 1: 328-31 (original in GB-Ob MS MDM d. 13, fol. 103-04).

[15] Ferdinand Hiller, Felix Mendelssohn-Bartholdy: Briefe und Erinnerungen (Cologne: DuMont-Schauberg, 1874), 17; Charlotte Moscheles, trans. A. D. Coleridge, Recent Music and Musicians as Described in the Diaries and Correspondence of Ignatz Moscheles (New York: Henry Holt, 1873), 178.

[16] Letter from Mendelssohn to his family dated 11 May 1832 (Mendelssohn family letters in the New York Public Library [hereafter, "US-NYp"], no. 153; see Rudolf Elvers, ed., Felix Mendelssohn Bartholdy: Briefe (Frankfurt am Main: Fischer Taschenbuch, 1984), 160, 158.

[17] "Ouv. fertig." Bodleian Library, Oxford, MS M. Deneke Mendelssohn g. 3, fol. 2r.

[18] Moscheles states that this exchange occurred on Sunday, 1 May (Recent Music and Musicians, 178-79); however, in 1832 that date fell on Tuesday, not Sunday. Since the first Sunday in May fell on the 6^th, that is most likely the date for the exchange.

[19] See Todd, Mendelssohn: "The Hebrides," 34.

[20] British Library, London, Add. MS 33465, fol. 251.

[21] Bodleian Library, Oxford, MS M. Deneke Mendelssohn g. 3, fol. 4v.

[22] See Wm. A. Little, "Mendelssohn and the Berlin Singakademie: The Composer at the Crossroads," in Mendelssohn and His World, ed. R. Larry Todd (Princeton: Princeton University Press, 1991), 65-85; further, Wolfgang Dinglinger, "Felix Mendelssohn Bartholdys Berliner Intermezzo, Juni 1832 bis April 1833," Mendelssohn-Studien 13 (2003): 101-23.

[23] See Dinglinger, "Felix Mendelssohn Bartholdys Berliner Intermezzo," 116.

[24] "Der Fehler der Composition ist vielleicht nur der, daß sie eines Commentars bedarf," Quoted from Eichhorn, Die Hebriden, 68.

[25] [Leipzig] Allgemeine Musikalische Zeitung (hereafter, LAMZ) 35 no. 12 (20 March 1833): 196: "Die Ouvertüre zu den ‘Hebriden’ von F. Mendelssohn-Bartholdy war für dieses Concertpublicum zu ernst und dürfte auch in sich selbst weniger abgeschlossen sein, als die Ouvertüre zum ‘Sommernachtstraum.’"

[26] Letter from Mendelssohn to Breitkopf & Härtel dated 9 August 1833, quoted in Rudolf Elvers, ed., Felix Mendelssohn Bartholdy: Briefe an deutsche Verleger (Berlin: Walter de Gruyter, 1968), 29.

[27] Letter from Mendelssohn to Breitkopf & Härtel dated 18 September 1833, quoted in Elvers, Verleger, 30-31.

[28] Letter from Breitkopf & Härtel to Mendelssohn dated 27 September 1833, held in the Bodleian Library, Oxford, GB II, 111. For a summary of this letter see Elvers, Verleger, 30 n.4.

[29] Letter from Breitkopf & Härtel to Mendelssohn dated 4 October 1833, held in Bodleian Library, Oxford, GB II, 117. For a summary of this letter see Elvers, Verleger, 30-31 n.4.

[30] See Elvers, Verleger, 31 n.1.

[31] A reference to the composer’s letter of 18 September 1833 (see n.29). As will be shown later, the German edition of the piano-duet arrangement carried a French title.

[32] Letter from Mendelssohn to Breitkopf & Härtel dated 29 November 1833, quoted in Elvers, Verleger, 31-33: ". . . sende heut mit der fahrenden Post die Partitur meiner Ouvertüre: die Hebriden und bedaure von Berlin aus erfahren zu haben, daß man Ihnen die ausgeschriebnen Stimmen nicht auf Ihren Wunsch schicken konnte. Man glaubte dort, ich hätte sie hier bei mir; und so fürchte ich, daß sie verloren sein werden. Auf jeden Fall aber ersuche ich Sie, den Stich der Stimmen nach dieser Partitur besorgen zu lassen, in der ich noch mehreres geändert und alle Zeichen genau gesetzt habe; auch bitte ich Sie den Stecher besondre Genauigkeit zu empfehlen für die p. und f., Crescendo’s, &c., auch für die oben angemerkten Buchstaben zum Wiederanfangen, die in allen Stimmen stehen müssen: Cello und Baß brauchen natürlich nicht durchgängig 2 Systeme zu haben (wie in der Partitur) sondern dies muß nur dann eintreten, wenn das Cello vom Contrabaß abweicht, sonst nicht. Es wäre mir lieb, wenn Sie mir eine Correctur dieser Stimmen zuschicken könnten, da ich dies Stück gern recht correct erscheinen sähe.

"Zugleich erlaube ich mir eine Anfrage. Ich war in Ihrer musikalischen Zeitung von einiger Zeit sehr streng getadelt, daß ich nicht die Partitur meiner Ouvertüre zum Sommernachtstraum herausgäbe, man hatte dies mir zur Last gelegt, und als eine Furcht vor der Kritik gedeutet. Nun ist es aber gerade im Gegentheil von jeher ein Lieblingswunsch von mir gewesen, einige meiner Partituren, die ich selbst lieb habe, dem Publikum vorzulegen, weil ich glaube, daß sie meinem Namen keinen Nachtheil bringen würden. Ich hielt es nur bisher für unmöglich und würde auch jetzt Ihnen nicht davon schreiben, wenn nicht einerseits der erwähnte Aufsatz und dann noch andre Gründe mich glauben ließen, daß eine solche Publikation jetzt vielleicht zu bewerkstelligen sey. Ich wollte also Sie fragen ob Sie wohl drei Ouvertüren die zum Sommernachtstraum, die Hebriden, und eine dritte in derselben Art in Partitur herausgeben könnten? Es müßte eine Opuszahl bekommen, und würde lange nicht so bogenreich, wie eine Beethovensche Symphonie, könnte also ziemlich wohlfeil werden. Auf Honorar würde ich natürlich bei solchem Unternehmen Verzicht leisten, und mich nur freuen, meinen Wunsch erfüllt zu sehen. Ich würde Ihnen nicht diese Proposition gemacht haben, da Sie in der letzten Zeit schon soviele Sachen nacheinander von mir herausgegeben haben, aber da zwei dieser Ouvertüren Ihr Eigenthum sind so konnte ich nicht umhin Sie zuerst darum zu befragen.

"P.S. Bitte, geben sie [sic] der Ouvert. zu den Hebriden einen Deutschen Titel, sowie ich ihn damals aufschrieb. Die Französischen sind nun einmal meine bêtes noires."

[33] See Alfred Dörffel, Geschichte der Gewandhausconcerte zu Leipzig (Leipzig: Breitkopf & Härtel, 1884), 2, "Statistik": 38.

[34] LAMZ 36 no. 13 (26 March 1834): 208.

[35] See the third Intelligenz-Blatt to the LAMZ for 1834, published with the issue for 5 March 1834 [LAMZ 36/10]. The piano-duet arrangement, which had been released in October 1833, is described here as "so eben erschienen," while the piano-solo arrangement is "unter der Presse."

[36] Elvers, Verleger, 33-34.

[37] Letter from Breitkopf & Härtel to Mendelssohn dated 21 April 1834, held in Bodleian Library, Oxford, GB III, 153. For a summary of this letter see Elvers, Verleger, 36 n.2.

[38] Dörffel, Geschichte, 2, "Chronik," 38.

[39] The Intelligenz-Blatt of LAMZ advertises the parts as being available for the 1834 publishers’ fair for Easter, which in 1834 fell on 30 March. The review is found at LAMZ 36 no. 26 (25 June 1834): 428f.

[40] Letter from Mendelssohn to Breitkopf & Härtel dated 29 July 1834, quoted in Elvers, Verleger, 36: “Hiebei erfolgen die Correcturen der beiden Ouvertüren. Ich habe hie und da noch einige Aenderungen machen müssen und bitte Sie daher den Stecher anzuweisen alle meine Bemerkungen genau zu befolgen. Dann wäre es mir lieb wenn die drei Ouvertüren nun recht bald erscheinen könnten, auch schon der Dedication wegen, um die ich schon seit langer Zeit angefragt und Antwort bekommen habe.”

[41] Letter from Breitkopf & Härtel to Mendelssohn dated 8 August 1834, held in the Bodleian Library, Oxford, GB III, 240. For a summary of this letter see Elvers, Verleger, 38 n.1.

[42] Letter from Mendelssohn to Breitkopf & Härtel dated 14 August 1834, quoted in Elvers, Verleger, 38.

[43] See Elvers, Verleger, 40.

[44] See Elvers, Verleger, 40.

[45] See Elvers, Verleger, 41-42.

[46] Again, the publisher’s response is lost, but its existence is documented by a stamp on Mendelssohn’s letter of 16 January. See Elvers, Verleger, 43 n.1.

[47] See Elvers, Verleger, 44-45.

[48] Mendelssohn’s older sister, Fanny Hensel, wrote to him on 19 March 1835 that she had received the scores for the three overtures; see Marcia J. Citron, ed. and trans., The Letters of Fanny Hensel to Felix Mendelssohn ([Stuyvesant, New York]:, Pendragon, 1985), 181, 495. The release is advertised in the Intelligenz-Blatt of the LAMZ on 1 April 1835.

[49] See Elvers, Verleger, 45.

[50] See Todd, Mendelssohn: "The Hebrides," 27-29.

[51] For an inventory of the contents of MS M. Deneke Mendelssohn c. 47 and a discussion of its provenance, see Margaret Crum, Catalogue of the Mendelssohn Papers in the Bodleian Library, Oxford, Vol. II: Music and Papers, Musikbibliographische Arbeiten, 8 (Tutzing: Hans Schneider, 1983), 13-14.

[52] For an explanation of this dating, together with careful analytical commentary, a facsimile, and a diplomatic transcription, see Todd, "Of Sea Gulls and Counterpoint," 197-213.

[53] For valuable assistance in various aspects of my investigation of AR I wish to thank J. Rigbie Turner, Mary Flagler Cary Curator of Music Manuscripts and Books of the Pierpont Morgan Library, New York.

[54] See Georg Eineder, The Ancient Paper-mills of the Former Austro-Hungarian Empire and Their Watermarks (Holland: The Paper Publications Society, 1960), plate 399.

[55] For information pertaining to the structure and foliation of AR I owe a debt of gratitude to Ms. Karen Desmond.

[56] For a discussion of this curious omission and Gounod’s comment, see Todd, "Of Sea Gulls and Counterpoint," 202. There is one further likely reason for Mendelssohn’s omission of the half-note D in the contrabass at this point: the same pitch in the same octave occurs in the moving line of the cello in that measure; a half-note D would have interfered with this cello line.

[57] In fact, Moscheles miscounted the number of measures in the printed score. The passage in question actually comprises twenty-six rather than twenty-four bars.

[58] For detailed information concerning source APf I am especially grateful to Peter Ward Jones, Head of the Music Section of the Bodleian Library, Oxford.

[59] On Mendelssohn’s relationship with the Horsley family, see especially Rosamund Gotch, Mendelssohn and His Friends in Kensington: Letters from Fanny and Sophy Horsley Written 1833-36 (London: Oxford University Press, 1936); further, Brigitte Richter, Frauen um Felix Mendelssohn Bartholdy, in Texten und Bildern vorgestellt (Frankfurt am Main: Insel, 1997).

[60] For a useful preliminary discussion of these differences, see Ward Jones, "Mendelssohn Scores," 68-69.

[61] See Peter Ward Jones, Catalogue of the Mendelssohn Papers in the Bodleian Library, Oxford, Vol. III: Printed Music and Books (Tutzing: Hans Schneider, 1989), 308.

[62] In the case of AL, too, I am thankful to Peter Ward Jones for his useful ideas and his many patient and insightful responses to queries.

[63] Ernest Walker, "Mendelssohn’s Die einsame Insel." Music and Letters 26 (1945): 148-50; Gerald Abraham, "The Scores of Mendelssohn’s Hebrides," Monthly Musical Record 78 (1948): 172-76.

[64] See, in addition to the catalogues by Margaret Crum and Peter Ward Jones already cited, Margaret Crum, Catalogue of the Mendelssohn Papers in the Bodleian Library, Oxford, Vol. I: Correspondence of Felix Mendelssohn Bartholdy and Others (Tutzing: Hans Schneider, 1980).

[65] The flyleaf, also in oblong format, bears a watermark including the initials "JFN" (information kindly provided by Peter Ward Jones).

[66] Beginning in the early 1830s, Mendelssohn typically inscribed the invocation "H. D. m." ("Hilf Du mir") at the head of the first page when drafting a new work or beginning a new manuscript for a composition already worked on elsewhere. The invocation "L. e. g. G." (present in AR) falls out of use around this time.

[67] Ward Jones, "Mendelssohn Scores," 69-70.

[68] Margaret Crum, Catalogue Vol. II, 19.

[69] For permission to study and discuss source CD, and for assistance in procuring a reproduction of that manuscript, I am indebted to Dr. Karl Wilhelm Geck and the Musikabteilung of the Sächsische Landesbibliothek – Staats- und Universitätsbibliothek, Dresden.

[70] See Felix Mendelssohn Bartholdy, Die Hebriden Overture, Op. 26, ed. Christopher Hogwood (Kassel: Bärenreiter, 2004).

[71] See Eichhorn, Die Hebriden, 15.

[72] See Ward Jones, "Mendelssohn Scores," 67-68.

[73] The most obvious such evidence is the false entry of the violas two measures early for seven measures beginning in mm. 94-100 (fol. 10v-11v). See Ward Jones, "Mendelssohn Scores," 69-70.

[74] Thanks are due to Dr. Martina Rebmann of the Musikabteilung of the Badische Landesbibliothek, Karlsruhe, for assistance in procuring a reproduction of source CK.

[75] The editions discussed below are only those of historical significance in the Overture’s textual dissemination. Other editions are identified in Bärbel Pelker, Die deutsche Konzertouvertüre (1825-1865): Werkkatalog und Rezeptionsdokumente, 2 vols. (Frankfurt am Main: Peter Lang, 1993), 1: 504-14.

[76] Letter of 29 July 1834 (see above).

[77] This situation is exemplified by the string parts. Violin 1 includes rehearsal letters A through C but omits the remaining three letters specified by Mendelssohn (D through F). Violin 2 includes A through D and omits the remaining two. The viola part includes A, B, and D but omits C, E, and F; and the Cello/Bass part includes A through C as well as E and F, but omits D.

[78] The exemplar of ES examined for this study is that held in the Staatsbibliothek zu Berlin – Preussischer Kulturbesitz. I am indebted to Dr. Hans-Günter Klein and Dr. Hellmundt Hell for their assistance in procuring a microfilm copy of this source.

[79] See Elvers, Verleger, 45 n.1.

[80] See note 48 above.

[81] On the general instances of this problem in Mendelssohn’s oeuvre, see Ralf Wehner, "‘It seems to have been lost’: On Missing and Recovered Mendelssohn Sources," in The Mendelssohns: Their Music in History, ed. John Michael Cooper and Julie D. Prandi (Oxford: Oxford University Press, 2002), 3-25.

[82] As noted above, the composer returned the corrected proofs for the orchestral score to Breitkopf & Härtel on 15 November 1834.

[83] See Elvers, Verleger, 30-31.

[84] See Ward Jones, "Mendelssohn Scores," 68.

[85] Because of the complexity and contradictions among the Hebrides Overture’s manuscript sources, a few words about editorial method are necessary. To begin with, the varying needs of the discussion for each example necessitates the identification of the precise copy-text used for each example, the approximate date of that text, and the correctional status of variants transmitted in the passage transcribed. Whenever the example gives the ante correcturam text, the edition indicates this via an "[a.c.]" at approximately that point in the measures affected; post correcturam readings are denoted by a "[p.c.]" in the appropriate points. Moreover, because Mendelssohn’s compositional shorthands often are suggestive of his ideas on phrasing or articulation, the examples indicate all groupings of notes by treating them as ligatures (i.e., by means of square brackets placed above or below the notes grouped in the autograph). Finally, all editorial interventions are graphically signified by the use of different fonts (for dynamic and expressive directions), perforation (for slurs absent from the copy text but presumably intended), and ficta (for accidentals omitted but presumably intended).

[86] Todd, Mendelssohn: "The Hebrides," 31.

[87] Todd documents that this subject evolved over the course of the Overture’s protracted compositional history, from Mull and Morven to Fingal’s Cave and Ossian. See Todd, "The Hebrides" and Other Overtures, 26-37, 78-83.

[88] Todd, "The Hebrides" and Other Overtures, 31-33.

[89] For a reproduction and transcription of the sketch and a discussion of its dating see Todd, "Of Sea Gulls and Counterpoint," 204-13.

[90] Ward Jones, "Mendelssohn Scores," 69.

[91] Of course, in addition to the classes of variants described below there are also mistakes – most often, printer’s errors. For a discussion of these issues, see Pietro Zappalà, "Editorial Problems in Mendelssohn’s Organ Preludes, Op. 37," in The Mendelssohns: Their Music in History, 27-42.

[92] See Eichhorn, Die Hebriden, 33-43.

[93] As is well known, serious philological explorations of nineteenth-century repertoires stem principally from the work of Gustav Nottebohm, whose pioneering work on the Beethoven sketchbooks also launched serious source-critical scholarship concerning the works of many other nineteenth-century composers in the early and mid-twentieth century. In Mendelssohn’s case, the first extensive study of compositional processes had to await the sesquicentennial of his birth. The landmark study in this line of inquiry was Donald Mintz’s dissertation (Donald Monturean Mintz,"The Sketches and Drafts of Three of Felix Mendelssohn’s Major Works," 2 vols. [Ph.D. diss., Cornell University, 1960]). Mintz’s initiative was taken up by numerous scholars beginning the mid-1970s; for an inventory of the most significant contributions, see Ch. 6 of my Felix Mendelssohn Bartholdy: A Guide to Research (New York: Routledge, 2001), esp. 211-16.

[94] Recent scholarship has identified a number of compositions that need to be added to traditional inventories of Mendelssohn’s works, as well as numerous other instances in which the texts by which the works are conventionally known are seriously flawed. For an overview of these findings, see my "Knowing Mendelssohn," Notes 61 (2004): 35-95.

[95] Mendelssohn Bartholdy, Die Hebriden, ed. Christopher Hogwood (Kassel: Bärenreiter, 2004).