The problems of tempo in Puccini's soprano arias are surprisingly vexing for performers, given that the composer provided many indications in his scores, including many metronome markings, and supervised the preparation of several singers who went on to make early phonograph recordings of his arias. The difficulties arise from the lack of markings in some cases, ambiguous or impractical markings in others (with some evidence that at times Puccini himself was not reliable in this matter), doubts about the authorship of some markings, and wide variations in tempo among recorded performances.
Tempo problems in Puccini's arias are far too many and complex to be treated exhaustively in a single article.1 Because of my own voice type and performing experiences, I have chosen an initial investigation of these problems with regard to the soprano arias in La Bohème, Tosca, and Madama Butterfly.
Three categories of information exist concerning Puccini's intentions with respect to tempo in these soprano arias: metronome markings, tempo words, and modification of tempo in Puccini's autographs, as well as in the first editions of his piano-vocal scores published in Milan by Ricordi. Determining Puccini's intentions on the basis of these sources requires care since, as the noted Puccini scholar Linda B. Fairtile has observed, "Puccini revised each of his completed operas at least once after its first performance, and you can often find different tempo markings in subsequent editions of the same opera. It's clear that Puccini was sometimes influenced by performers in this point."2
In 1954, Luigi Ricci, who had been Puccini's rehearsal pianist, published a guide and commentary on the performance of Puccini's operas, entitled Puccini interprete di se stesso ("Puccini, Interpreter of Himself"), which has been translated by Harry Nicholas Dunstan, Jr., under the title Performance Practice in the Music of Giacomo Puccini as Observed by Luigi Ricci (Ann Arbor: UMI, 1989). While there are certainly many details and insights in Ricci's book that go beyond what can be found in Puccini's scores, he still leaves a great many more questions unasked and unanswered, including many that relate to tempo. A generous sample of these questions is addressed in the paper that follows, with their frequent references to Dunstan's translation of Ricci's book.
I have observed two major phenomena in the early recordings and the scores of Puccini's soprano arias in La Bohème, Tosca, and Butterfly: First, that Puccini's initial tempo markings, even when indicated by metronome settings, often differ from the tempo actually used by early performers. Second, the changes of tempo during the course of an aria are interpreted quite differently from one early recording to another. Some of Puccini's indications have never been followed as far as I can tell. The different treatment of rubato, fermata, portamento, dynamic range, expression markings, etc. in Puccini's arias all affect the tempo of an aria as sung by singers with different vocal qualities and in different time periods.
There are even reasons to doubt that Puccini always carefully considered his own metronome markings. Giulio Ricordi wrote to Puccini on November 29, 1895, two months before La Bohème's first performance:
Concerning the metronomic markings. . . . Have you been able to try them out well at the pianoforte? There are two that seem (too) fast to me, thinking of the orchestral rhythm and of the pronunciation of the syllables. Try (these): Act 2 [quarter = 126—all the notes two to a beat seem to me difficult to pronounce at this tempo: 'Falso questo re'—and so on. Chorus of boys [quarter = 132]—it seems to me difficult to pronounce and to understand the syllables 'Col suo bel carretto tutto lumi e fior!' Write to me at once if you (wish to) confirm the tempo indicated, because I must prepare the parts, and afterwards I cannot make further corrections of any kind.3
To obtain a better understanding of this letter, I checked the metronome marking in Puccini's first edition (plate number 99000). The metronome setting in the score is quarter = 112 instead of the 126 mentioned by Ricordi. Evidently Puccini changed the tempo marking in response to Ricordi's letter. Even so, the music of the "Falso questo re" still appears to be difficult for the baritone singing the role of Schaunard, not only because of the very quick pronunciation of the words but also because of the leaping notes in the high vocal range with a strong dynamic requirement. On the other hand, the score maintains Puccini's original tempo (quarter = 132) for the Chorus of Boys. In the recording of La Bohème (NI 7862/3) as performed at La Scala and conducted by Umberto Berrettoni in 1938, the conductor gave a tempo of quarter = 123 to the Chorus of Boys in Act II. Significantly, the syllables of the text were still not pronounced clearly.
The ambiguities of base tempo and tempo change can also be found in Musetta's aria "Quando me'n vo." The metronome marking at the beginning of Musetta's waltz, as contained in Puccini's first edition, is quarter = 104 with the tempo words Tempo Di Valzer Lento and expression marks Con molta grazia ed eleganza. I have found no performer who chose this metronome number; in fact, many early performers started the tempo much slower, for example quarter = 85. Puccini noted tempo modifications such as a tempo and quasi rit, a tempo, and appena allarg for each phrase of the aria. In recordings of the early period, there was only a slight contrast, but over time, the tempo slowed in recorded performances and one hears a large contrast between the rit and a tempo. What is the intended margin of tempo modification in this aria, and how does it correspond to the metronome number and the tempo words at the beginning of the number? The ambiguity of tempo results in different interpretations.
At the beginning of the aria "Si, mi chiamano Mimi," Puccini's metronome setting is quarter = 40, but the duration of each performance in the many recordings to which I have listened differs significantly from that, even though their tempos are initially close to Puccini's metronome setting. At rehearsal number 37 of this aria, Puccini wrote quarter = 144, a faster tempo with the expression mark con semplicità, which conflicts with his tempo words All.tto Moderato. Just which tempo did Puccini intend? Is it the metronome number or the tempo words?
In his book Puccini interprete di se stesso (1954), Luigi Ricci, Puccini's rehearsal pianist, carefully recorded the durations of individual acts performed under Puccini's supervision, as well as all of Puccini's comments about his works from rehearsals, (including tempo changes and modifications of tempo) and comments related to the tempos when they were not exactly as they appear in Puccini's original scores. Ricci's book, however, is still not sufficient for determining the intended tempos in Puccini's soprano arias. In his "Ten rules of Puccini," Ricci wrote:
Puccini had a very acute sensibility regarding the variety of rhythmic movement; as much for the fundamental or general indications as for those restrictive or secondary ones. Regarding andante, for example, he wanted that in reality it would correspond to its exact movement which is moderate, rather slow, but not too slow.4
Here, Ricci also mentioned several of Puccini's disagreements regarding the slow tempos during rehearsals such as "too slow" and " Life, life Maestro! Don't slow down too much."5 What is the standard for adjusting the slower tempo with which Puccini was not satisfied during rehearsals, and why did these situations frequently arise? Was Puccini's tempo misunderstood by conductors, or was there a disparity between Puccini's imagined tempo and the actual performance tempo he preferred? How is the modification of tempo applied in different arias? There are no apparent answers. Puccini's own term, "equilibrium," which "presides over all his poetry, over all his aesthetics, over all his art of composition,"6 could aid performers in understanding Puccini's style of music, but it also leaves much room, maybe too much, for performers to interpret his tempo freely.
Concerning Puccini's metronome tempo, Ricci indicated in his book:
The metronome numbers written by Puccini in the piano/vocal score are very exact. Some numeric indications have been changed, some others increased. There is not a metronome sign which has not been scrupulously checked by me in each of the many performances in which I assisted the author. But one must understand Puccini clearly regarding the metronome.7
And this raises yet another question: If Puccini wrote the exact metronome number in his vocal score, which edition represents Puccini's ideal metronome setting? If the metronome setting was correctly written on the vocal score by Puccini, why were the changes or increases of the metronome setting still required during rehearsals? For instance, in the fifth measure of the aria "Si, mi chiamano Mimi," Puccini did not write any tempo marking on his vocal score (first edition, plate number 99000), but Ricci pointed out in his book, "At bar 5 of page 70, the tempo is un poco più mosso, and the M.M. is precisely quarter = 46—based on repeated inspections made with the Maestro."8 Performers could be confused by Puccini's original scores and Ricci's commentary.
When commenting on Puccini's ten rules, Ricci indicated, "Puccini wanted the notes with a 'bird's eye' in the middle of a phrase to be exactly double their morphological value."9 However, if one carefully looks at the tempo markings on Puccini's scores, the value of Puccini's fermatas with his modifications of tempo occasionally does not exactly correspond to his rule, even if "Puccini demanded exactness in performance of that which was written."10 For instance, in the last phrase of Butterfly's final aria, "Guarda ben! Amore, addio! addio! piccolo amor!," Puccini indicated the markings rit. both at the beginning "guarda" and the final "piccolo amor." In addition, he placed a fermata on the "-mo-" of the first "amore" in the middle of the phrase. The value of the fermata, as interpreted by different performers, will vary based on the different tempo of the marks rit. The same situation occurs in Mimi's aria "Si, mi chiamano Mimi," measures 52-55. Here Puccini wrote poco allarg. on high A. Ricci stated, "Puccini did not want any corona on the A: not even the smallest fermata."11 But in fact, the value of the A with an accent mark has to vary based on the value of poco allarg. The rule regarding Puccini's fermata as presented by Ricci, is not clear and is sometimes even inconsistent with Puccini's own tempo markings in the original scores.
In addition, Ricci pointed out in his book, "Puccini disliked the portamento at any cost, the portamento at each turn, the habitual portamento. But if by chance he marked one, woe to whomever did not give maximum expressiveness to the desired piece."12 In Puccini's scores, especially in his soprano arias, there are many portamentos. In the aria "Si, mi chiamano Mimi," as printed in the first edition of the opera, eleven portamentos appear. Puccini frequently uses the "portamento" indication in those soprano arias which are particularly expressive emotionally. Singers from different time periods supply their own additional tempo and usually interpret Puccini's portamentos with a slowing of tempo. In the first measure of this aria, Ricci also noted that "the legatura is observed with a portamento della voce between the and the E."13 Puccini's concept of portamento mentioned in Ricci's book is thus not confirmed in Puccini's scores.
The most striking difference between Puccini's autograph and the first edition of the vocal scores is that in his autograph Puccini did not indicate a metronome setting for each of his early arias. From where then did the metronome setting in the first editions come? No revision of metronome indications has yet come to light. There is, however, an example from Mimi's first aria. Puccini wrote tenuto on the second beat of measure 23 in his autograph, rather than the fermata that is found in the first edition of the vocal score. Ambiguity of tempo is again the result.
Arturo Toscanini, the first conductor of La Bohème, and Sir Thomas Beecham, who conducted La Bohème more than 300 times, both conducted Puccini's operas in the composer's presence, and both recorded complete performances of La Bohème. However, neither Toscanini nor Beecham's recordings of Puccini's operas and arias are definitive as regards tempo because discrepancies in tempo exist not only between them but also among their own recordings. Puccini's metronome numbers, tempo words, and modifications interpreted by the two conductors vary with the singers and over different periods of time. Although sometimes the total duration of a recorded performance of an aria under the direction of one conductor approximates another performance of the same aria in a different year, the tempos of these two performances can vary considerably, even though the conductor remains the same. Toscanini's tempo was usually faster and Beecham's slower. Neither conductor, however, consistently adhered to Puccini's tempo as indicated in the first editions.
Nor are the recordings made by singers of Puccini's time, even those singers who sang under the direction of the composer himself, adequate for answering the questions about Puccini's intended tempos, because, in the matter of tempo, the early recordings are by no means in agreement with one another. In fact, they vary considerably among one another and with Puccini's published markings.
One striking example is the four markings on high As in Mimi's first aria. The first A is marked with ritardando, the second with con grande espansione, the third with poco allargando, and the fourth with tenuto. I have not found these four marked As treated similarly in any early recordings. Singers seem to prefer spending a longer amount of time on those high As in order to display their singing skills and expressive emotions. Some spinto or dramatic sopranos treated the tempos of the four As more freely than did lyric sopranos, and the phrases were sometimes broken because of their need to take a breath.
In 1932, a review of Lily Pons's recording appeared that included Mimi's first aria (Parlo. R20204, 12in., 6s). In it, her tempo was described as "somewhat more sober than we are accustomed to now, but it is actually about the same prime donne employed in the years gone by, when operatic life was less strenuous."14 Just what, however, is the standard Puccini tempo against which Lily Pons' performance was judged? The differences in various singers' phrase shaping, portamento, dynamic range, and expression marks, their different vocal qualities, and even their social positions all influence the arias' tempo.
The period in which a recording was made also influences the recorded tempo of each aria. The flexible tempo was dependent upon the size of each recording and the speed required by different companies' recording equipment. There were non-standard rpm requirements in early recordings (labeled as "78 rpm") which produced some unusual acoustical phenomena. For example, the tempo in Melba's 1910 recording of "Vissi d'arte" (LRX 2821) is slower (quarter = 34) than the indication (quarter = 40) on the vocal score, and the duration of the aria is 3'15", a time longer than the duration of other recordings of the aria made during the same time period. Her pitch is also lower than normal. Has the slower playback speed perhaps changed the tempo and made the pitch a half step lower?
In addition, the tempo may have been influenced by a need to hurry the performance so that it would fit onto a single side of a record. For instance, there is no recitative in Mimi's final aria sung and recorded by Florence Easton in 1928 (LT-10'2257) or by Delia Reinhardt, whose performance of the same aria was recorded between 1922-1929 (LV-1421). The average duration of "Vissi d'arte" is around 3-4 minutes, but in Perras Maria Gerhart's recording (LV-73) made between 1935-37, the aria lasts a mere 2'09". This does not, however, explain differences of tempo in early recordings of one and the same aria on similarly sized records with the same capacity. It also does not explain the presence of unusually slow tempos, which are found frequently in the early recordings. Please see the following table as an example of tempo discrepancies in the aria "Vissi d'arte, vissi d'amore."
|Vissi d'arte, vissi d'amore|
|m.m.1||m.m. 14||m.m. 22||m.m. 26|
|score And.te Lento||q = 40|
|Ricci||q = 40||a tempo||a tempo||year|
By examining Puccini's autographs, the early editions, and many early recordings, and additional sources of information such as Luigi Ricci's commentaries and various personal letters and references, I found that the most successful interpretations of Puccini's tempos in the early recordings of his soprano arias occur in Butterfly's last aria, "Con onor muoreTu, tu, piccolo Iddio," where singers can rely on Puccini's metronome settings to achieve the intended dramatic effect. These settings appear both in the vocal score and in Ricci's commentary. However, neither Puccini's vocal score nor Ricci's commentary provides sufficient tempo indications for the early performers of many other arias, especially those in La Bohème and Tosca, although in the later Madama Butterfly, the tempo indications seem to correspond more closely to actual recorded versions.
From La Bohème to Tosca and through to Madama Butterfly, Puccini's tempo markings and Ricci's commentary demonstrate a pattern of development. In the early compositions, many early singers, while generally adhering faithfully to Puccini's slower markings, failed to comply with his faster indications. This was also consistent across the spectrum of early recorded versions of these arias. By the time Puccini wrote Butterfly (1903), his tempo markings seem to be closer to how singers actually interpreted his music, and more reliable. Puccini's tempo indications after this point were modified by the singers only to a small degree. The range of metronome settings for these arias is smaller than in his earlier works.
The fact that Puccini's tempo markings in Butterfly's last aria are closely matched to those used by singers may be attributed to a variety of possible reasons. The first may be due to the fact that Puccini was beginning to understand the performers' abilities to convey the dramatic intent of the music and text. Secondly, he was also gaining experience and understanding of how tempo affects dramatic intent in performance. In addition, his continued association with performers of his operas certainly left him with a greater understanding and appreciation for tempo indications. And finally, he was simply maturing as a composer, gaining a better understanding of tempo that comes with experience.
Apel, Willi. The Harvard Dictionary of Music. Cambridge: Harvard University Press, 1956.
Ashbrook, William. The Operas of Puccini. Ithaca: Cornell University Press, 1985.
Clough, Francis F. and Cuming, G. J. The World's Encyclopaedia of Recorded Music. Westport, Conn.: Greenwood Press, 1970.
Fairtile, Linda Beard. Giacomo Puccini's Operatic Revisions as Manifestations of his Compositional Priorities. Ann Arbor, MI: UMI Press, 1996.
Fairtile, Linda Beard. Personal letter, July 26, 1997.
Klein, Herman. Herman Klein and The Gramophone, Being a Series of Essays on the Del Canto (1923), the Gramophone and the Singer (1924-1934), and Reviews of New Classical Vocal Recordings (1925-1934), and Other Writings from the Gramophone by Herman Klein, ed. William R. Moran. Portland: Amadeus Press, 1990.
Kutsch, K. J. and Leo Riemens. A Concise Biographical Dictionary of Singers, trans. Hany Earl Jones. NY: Chilton Book Company, 1962.
Pleasants. Henry. The Great Singers, from the Dawn of Opera to Our Own Time. New York: Simon Schuster, 1966.
Puccini, Giacomo. La Boheme, vocal score, plate number 99000. Milan: Ricordi, ca. 1896.
Puccini, Giacomo. Madama Butterfly, vocal score, plate number 110000. Milan: Ricordi, ca. 1907.
Puccini, Giacomo. Tosca, English version, "Critical Commentary," ed. Roger Parker. Milan: Ricordi, 1996. Pp. xviii-xxxviii.
Puccini, Giacomo. Tosca, vocal score, plate number 103050. Milan: Ricordi, ca. 1899.
Puccini Librettos, trans. William Weaver. New York: Doubleday & Company, Inc. 1966.
Ricci, Luigi. Performance Practice in the Music of Giacomo Puccini as Observed by Luigi Ricci, trans. Harry Nicholas Dunstan, Jr. Ann Arbor, MI: UMI Press, 1989.
Ricordi, Giulio. Letter written to Giacomo Puccini on letterhead printed "G. Ricordi & C. Milano," Milan, November 29, 1895.
Sadie, Stanley, ed. The New Grove Book of Opera. New York: St. Martin's Press, 1997.
Singher, Martial. An Interpretive Guide to Operatic Arias. Pennsylvania State University Press, 1983.
The Victor Book of the Opera, third revised edition. Camden, NJ: Victor Talking Machine Company, 1912.
Walker, Malcolm, comp. Opera on Record, ed: Alan Blyth. London: Hutchinson & Co. 1979.
Wilson, Conrad. Giacomo Puccini. London: Phaidon Press Limited, 1997.
Si, mi chiamano Mimi
Cesira Ferrani, 2914b, G&T 53281, 1902
Claudia Muzio, Matrix: 298ai, HMV, cat: 053264, 1911
Claudia Muzio, ODYSSEY, Y33793, 1921
Claudia Muzio, Matrix: B 6985 Columbia, Cat: LX 583, 1935
Delia Reinhardt, LV-142, 1922-1929
Gabrielle Ritter-Ciampi, Brunswick, International 90009
Geraldine Farrar, Matrix: C12742-1, Victor, Cat: 88413, 1912
Gemma Bellincioni, Tima 35, 1905
Grace Moore, Brunswick 50140, 1927
Maria Cebotari, Electrola DB 4415
Licia Albanese, Victrola, VICS-6019 (e), 1946
Licia Albanese, Cat:DB 3448/60, Matrix: 2BA 2362/87, 1938
Rosetta Pampanini, Columbia, Operatic series, No.5-B, 67718-D `Bx 417, 1937
Zinaida Jurjewskaja, LV-89, 1922-1925
Quando me'n vo
Alma Gluck, Victrola, 649-A, 10-inch
Anne McKnight, VICS-6019, 1946
Bidú Sayao, Columbia, 17515-D, CO 37918
Conchita Surpervia, Parlophone Odeon Series, #RO20180, 10-inch, 78rpm
GinaViafora, Acoustic, 64085, 1905-1906
Irene Pavloska, Brunswick, 13004-B, 3499B1
Luisa Bresonier, 2896b, G&T 53252, 1902
Lucine Amara, Seraphim, SIB-6000, (remastered, 1974), 1956
Maria Jeritza, LV-2, Odeon RXX 76537, Matr. Nr. XXB 6142, 1914
Maria Kurenko, Columbia, 2055-m, 143656, 1913
Tatiana Menotti, Cat: DB 3448/60, Matrix: 2BA 2362/87, 1938
Donde lieta uscì
Alma Gluck, Victrola, 649-B, 10-inch
Cesira Ferrani, 2915b G&T 53282, 1902
Claudia Muzio, Columbia, Cat: LB 40, Matrix: B 6958, 1935
Frances Alda Victor, Matrix: B 14487-1, 1914
Grace Moore, CA 7073-1, Brunswick 50140, 1927
Hina Spani, BK 2009-2, VRCS 1, 1926
Licia Albanese, Cat: OB 3448/60, Matrix: 2BA 2362/87, 1938
Licia Albanese, VICS-6019 (e), 1946
Lisa Perli, HQM 1234, 1936
Magda Oliviero, Archivio Italiano Cetra / LPO 2008
Nellie Melba, Angel Records, COLH 125, Master 28 assigned record: 03030, 1904
Selma Kurz, EMI, Historic Masters HMA 42, 1907
Victoria De Los Angeles, SIB-6000, 1956
Non la sospiri la nostra casetta
Geraldine Farrar, Matrix: C 9834-2, Victor, Cat: 88287, 1911
Vissi d'arte, vissi d'amore
Beita Kiurina, Polyphon, 50169, 1920
Claudia Muzio, Pathé, 63018, 1917
Emma Eames, Victor, 85059, 1905
Emma Carelli, Pathé, 4379, 1905
Emmy Destinn, Gramophone, 2-053053, 1911
Elda Cavaliere, Victor, 74054, 1906
Frances Alda, Victor, 74440, 1914
Geraldine Farrar, Victor, 88192, 1909
Hertha Stolzenberg, Horiochord 15982 ?, 1919
Helene Wildbrunn, Grammophon, 04335, 1919
Maria Jeritza, Odeon RXX 76538, 1914
Marie-Louise Edivina, HMV Gramophone 2-053150, 1919
Medea Mei-Figner, G&T 23134, 1901
Nellie Melba, Victor, 88075, 1907
Rosa Ponselle, Columbia, 5095-M, 1919
Salomea Kruszelnicka, G&T 23362, 1902
Selma Kurz, Gramophone, 2-053079, 1913
Ancora un passo
Geraldine Farrar, LCT 1, Victor, Form 2S4154, 1907
Geraldine Farrar, Matrix: B 4225-2, Victor, Cat: 87004, 1909
Geraldine Farrar, Mono, VIC-1600, 1916
Ieri son salita
Geraldine Farrar, Victor, 616-B, 1908
Geraldine Farrar, Mono, VIC-1600, 1909
Geraldine Farrar, Matrix: B 8267-2, Victor, Cat: 87004, 1909
Un bel dì vedremo
Amelita Galli-Curci, Victrola, #6130-B, 1921-1925
Berta Kiurina, LV-91, 1908-1930
Edith Mason, Brunswick, 50080, XE 16614, 1925
Elisabeth Rethberg, Brunswick, 50065, 1924
Elisabeth Rethberg, Odeon, 0-8387, 1930
Emmy Destinn, G&T; 053171, 0801v, 1908
Geraldine Farrar, Matrix: C 5055-2, Victor, Cat: 77113, 1909
Hina Spani, Columbia, (A.R), VRCS 1, 1926
Maria Gerhart, LV-75, 1924
Maria Jeritza, Odeon, RXX 76539, Matr, Nr, XXB6144, 1914
Natalie Ermolenko-Yuzhina, Historic Masters, HM 108, 1911
Rosa Raisa, Vocalion, #B 70036, 1920?
Rosa Ponselle, Matrix: 49571-2, Columbia, Cat: 68056D, 1919
Tamaki Miura, Columbia, 49260, 1917
Zinaida Jurjewskaya, LV-89, 1922-1925
Sai cos' ebbe cuore
Geraldine Farrar, Victrola, 617-A, 1908
Geraldine Farrar, Matrix: B 8269-1, Victor, Cat: 87055, 1909
Con onor muore. . . Tu, tu, piccolo Iddio
Emmy Destinn, G&T 7-53022, 12753u, 1908
Geraldine Farrar, Matrix: B 8270-1, Victor, Cat: 87030, 1909
Desmond Shawe-Taylor, A Biographical and Critical Sketch, (1961), COLH 125, program note, 2
Emily Coleman, Composers' Favorite Interpretations, RCA, Victor, LCT-1
Irving Kolodin, Beecham on "Bohème," Saturday Review for August 25,1956, IB-6000, program note
Francis Robinson, The Met's first "Butterfly," Mono VIC-1600, program note
"Notes from an interview with Licia Albanese in her New York Apartment in July 1994," NI 7862/3, program note, 19
Norman White, recording, NI 7814, program note, 1990
1My dissertation, "Tempo in Puccini's Soprano Arias," is available through UMI. I wish to express my sincerest gratitude to Professor John Walter Hill, who guided me through each stage of my writing. (University of Illinois, Urbana-Champaign, 1999).
2Letter from Linda B. Fairtile, July 26, 1997.
3Letter written on letterhead printed "G.Ricordi & C.Milano," November 29, 1895. I am grateful to Professor Gabriella Ravenni of the University of Pisa for locating and transcribing this letter for me.
4Performance practice in the music of Giacomo Puccini as observed by Luigi Ricci, trans: Harry Nicholas Dunstan, Jr. (UMI, 1989), 1.
5Performance practice, trans: Dunstan, 2.
6Performance practice, trans: Dunstan, 3.
7Performance practice, trans: Dunstan, 2.
8Performance practice, trans: Dunstan, 73.
9Performance practice, trans: Dunstan, 4.
10Performance practice, trans: Dunstan, 4.
11Performance practice, trans: Dunstan, 76.
12Performance practice, trans: Dunstan, 4.
13Performance practice, trans: Dunstan, 72.
14Herman Klein, Herman Klein and the Gramophone, edited by William R. Moran, (Portland: Amadeus Press, 1990), 545.