Contrasts and Comparisons from Baroque to Modern


Michelle Edelman, copyright 2001


  VML Graphic Rendering is necessary in order to view the graphics presented as part of this paper.  Your web browser should prompt you to download it if you do not already have it installed on your computer

When clicking on footnote reference numbers, after reading the footnote reference you can return to the footnote number by clicking on the footnote reference number in the section where the footnote reference has been placed.       


With the popularity of performance practice growing, thanks to people such as Wanda Landowska, Gustav Leonhardt, Eduard Melkus, the Dolmetsch and Kuijken families, and more recently violinists such as Monica Huggett, Andrew Manze, Ingrid Matthews, and many others, questions have begun to arise about the differences between modern day practices and historical performance practices.  The Devil’s Trill Sonata, by Tartini (1692-1770), provides an excellent example of these differences.  What makes the Devil’s Trill Sonata such an excellent example of the striking contrasts between baroque and modern violin performance is the fact that the best-known version of this sonata is not in its original form, but in an arrangement by Fritz Kreisler (1875-1962).  Although this version is perhaps better known than what we believe to be Tartini’s version, which was first published in 1798,[1] the supposedly authentic version of the Devil’s Trill Sonata is beginning to gain popularity. 

            Comparisons and contrasts between the two above-mentioned versions of Tartini’s Devil’s Trill Sonata provide an interesting study, considering that the original manuscript version was not Kreisler’s reference for his own arrangement of the work.  (It is not known what version of the sonata Kreisler may have used when he composed his own arrangement.)  The differences between Tartini’s and Kreisler’s versions of the sonata are due to the backgrounds of the versions of the sonatas, the contemporary bows and instruments in use, and the contemporary performance practices in music while each of the respective versions of the Devil’s Trill Sonata was being written/composed/arranged.

            The Devil’s Trill Sonata, by Tartini, has remained his most famous work.  It is interesting to note that Tartini claimed to have begun to write this sonata as the result of a dream.  Undoubtedly, the story that Tartini told involving this particular sonata has added to its mysticism and popularity.  J. J. de Lalande recants the story Tartini told to him.

          ”One night in 1713 he dreamed that he hade made a contract with the devil, who happened to be in his service.  Whatever Tartini wanted was granted to him, and all his wishes were anticipated by his new servant, who gave him his violin to see if he could play anything harmonious.  But what to Tartini’s surprise when he heard a sonata so original and lovely performed with such perfection and meaning that he could never have imagined anything like it!  He experienced such amazement, admiration and delight that he was breathless; this strong emotion woke him up and he immediately seized his violin in the hope that he would be able to remember at least part of what he had heard, but in vain.  The piece that Tartini composed then is indeed the best of all he has ever done, and he calls it The Devil’s Sonata.  But the former one that amazed him was so much higher that he would have broken his violin and given up music forever if only he could have.” [2]


            Many violin historians and scholars have raised the question as to when Tartini may have written his famous violin sonata.  J. J. de Lalande’s quote placing the events that lead to the Devil’s Trill Sonata in the year 1713 has been questioned because of the other inaccuracies present in Lalande’s works. [3]  Perhaps the best way to categorize Lalande’s account of the events leading up to the Devil’s Trill Sonata would be fanciful (with less than accurate information).  With these two thoughts in mind, many theories about when Tartini may have begun the sonata have been suggested.  One theory has stated that the sonata could not have appeared any earlier than 1730 or even as late as 1740.  However, another theory has also been advanced that shows itself to be in favor of Lalande’s account, Michelangelo Abbado has presented a theory revolving around the fact that Quantz discussed the Italian master’s skill in playing double trills, after hearing Tartini play in 1723.  Since the accompanied trill does not occur in any other of Tartini’s known works, Abbado is prone to think that the sonata was written before that time in 1723. [4]

            Tartini’s background leading up to the sonata provides a fantastic contrast to the events that led up to Kreisler writing his own arrangement of the Devil’s Trill Sonata nearly two hundred years later.  Kreisler was a wunderkind on the violin and had already won the gold medal at the Vienna Music Conservatory by the time that he was ten years old.[5]  In addition to studying to become a doctor and serving in World War I, Fritz Kreisler remained active on the violin.  It was around the year 1900 that Kreisler began to compose many of his short violin pieces that brought him additional fame.  These included Liebesleid, Liebesfreud, Praeludium and Allegro, and others too numerous to mention.  It was also during this time that Kreisler might have begun work on his own arrangement of the Devil’s Trill Sonata, which was published in 1905.[6]  Kreisler possessed a unique sound and style all of his own.  Part of his style consisted of expressive slides and position changes, and in many cases these traits would show through in any editions or arrangements that he would make as well.[7]

          Violins were still considered to be in baroque condition when Tartini wrote his famous Devil’s Trill Sonata.  This meant that both the violin and the bow that Tartini would have had in use at the time of would have been very different from the modern bow and violin that is now widely known and played universally.  The violin bow had been undergoing developments since its inception.  Tartini has been credited in helping to modify to the bow of his time by lengthening the bow overall and adding a more ‘modern’ square tip (also known as the ‘hatchet head’).  These changes to the bow could have been made around 1730.[8] 

          The classification that Robin Stowell uses to contrast the differences in bowing prior to the modernization of the bow and after its modernization uses Tourte’s revolutionary bow as a point of division.  According to this classification, the bows before Tourte are referred to as being pre-Tourte, and all bows after Tourte are referred to as Tourte.  With this classification in mind, it can be stated that the violin bow that Tartini is said to have helped develop is pre-Tourte. [9]  By Stowell’s classification of these bows into two categories, it also clarifies the musical strengths and weakness of both types of bow.  Pre-Tourte bows were not designed to produce an even tone, as the modern violin bow was, instead, pre-Tourte bows produced an uneven tone.  Sol Babitz’s theory about the basic bow stroke of the eighteenth century states that the basic bow stroke of the century was a crescendo-diminuendo stroke, resembling the messa di voce. [10]  Robin Stowell, who shows another example of common eighteenth century bowing, the most widely used stroke for the pre-Tourte bow was an articulated non-legato stroke, due to both the characteristics of the bow itself and the manner of the holding and drawing the bow. [11]  Both of the above mentioned bow strokes were considered the desired effect for the time, and as far as many were concerned, the bow needed no more improvement.  Below is a diagram illustrating the changes that took place in the violin bow from the time of Mersenne to Viotti, who, as legend has it, helped Tourte perfect the modern violin bow. (Figure 1)[12] 

1 Violin Bows 1620-1790 as illustrated by Fétis

The violin of the baroque era was very different from the modern violin.  Bridges on the violins were thicker and less curved in comparison with their modern counterparts, the angle of the neck and fingerboard in relationship to the instrument was in line with the instrument instead of sloping at an angle of twenty to thirty degrees away from the instrument, a graduated bass bar and belly, and a thinner sound post when compared with the modern violin.  Of course, there are other more obvious differences between the baroque and modern violin, such as the absence of a chin rest, the gut strings in use, and the significant differences between the pitch of A=415 MHz and around A=440 MHz.  With all of these differences in mind, it can be hard to imagine that the modern violin even resembles its early counterpart! 

Between the time of Tartini’s death and the birth of Fritz Kreisler, the violin had become fully modernized.  Acoustical demands on the violin had caused most of the changes from what is considered baroque to the present status of the violin.  Concert halls had changed dramatically since the time of Tartini when they were mostly used for small chamber ensembles, to the late eighteenth and early nineteenth centuries when they were used for large orchestras and opera productions.  In order to meet this new acoustical demand the violin was required to sound more brilliant and have a further carrying sound.[13]

 To accomplish this, the pitch was raised on the instrument from A=415 MHz to around A=440 MHz.  With the adjustment of the pitch came the rest of the changes to the violin.  In order to support the higher tension needed in the instrument for this more brilliant sounds, the violin needed to be physically reinforced and prepared to meet these new demands.  The violin bridge was thinned and raised, the bass bar was thickened, and the wood on the belly of the violin was regraduated.  Obvious changes to the violin included the new angle at which the fingerboard and neck were now sitting in comparison to the violin.  This change was necessary in order to physically reinforce the higher tension on the violin caused by the raising of the pitch.  (Interestingly, the regraduation of the wood on the belly actually weakened the violin, despite the many other efforts made to strengthen it physically.)  Had the angle of the neck and fingerboard on the violin not been changed, the violin would have collapsed because of the large increase in tension.  Examples of both of these changes to the violin are shown in figure two.[14]  By adjusting the fingerboard of the violin to match the angle of the strings, clearer fingering and better double stopping were now possible. [15]

Figure 2 On the left, the diagram shows the d typical 'baroque' bridge (top) compared with the common modern (bottom) variety.ifferences between the angles of the necks of the baroque (top) and modern (bottom) violins.  On the right, there is an outline of a



When Tartini wrote and applied the bowings to the Devil’s Trill Sonata, he was using a bow that was categorized as pre-Tourte and a violin that was still in baroque condition.  His bowings reflected the characteristics of the bow that he used to play the violin.  Many of the bowings that Tartini may have used would have used for this sonata would be considered inappropriate for the violinist of today hoping to achieve a similar effect with a modern violin and bow.  (see figures 5-10)  Therefore, when Fritz Kreisler did his edition of the Devil’s Trill Sonata he did what any typical romantic violinist would do.  He wrote bowings, fingerings, and dynamics appropriate to the time and audience for which he was writing for.  Many of the differences within the two versions of the sonatas deal directly with the physical changes to the violin because of the differences in the fingering or bowing then required of the violinist. 

Tartini wrote the Devil’s Trill Sonata during the culmination of the baroque style.  By 1713, when it is said that Tartini started work on the Devil’s Trill Sonata, there was already a highly developed pattern for the baroque violin sonata as laid down by Corelli through his sonatas.  This pattern of the sonata da chiesa or church sonata was that of slow-fast-slow-fast. [16]  In Tartini’s composition, the Devil’s Trill Sonata, he remained true to this pattern by writing four movements entitled Andante-Allegro-Andante-Allegro (titles of movements are taken from the Abbado edition based on manuscript versions of the sonata in the hands of Tartini’s students).  Kreisler also gave names to this pattern of slow-fast-slow-fast movements, Larghetto-Allegro energico-Grave-Allegro assai, which still reflect the baroque sonata da chiesa.  While the descriptions of these movements, as given by the composer, tell the musician the tempo and more importantly, the mood that they are to play these sections in, an important factor of baroque music is not included.  Most baroque music constituted some type of a dance or another familiar form.  The Andante, which begins the Devil’s Trill Sonata, is by definition, and examination, a sicilianoSiciliana, by definition, invoke a gentle pastoral mood, usually through 6/8 or 12/8 time.  Perhaps one of the best ways to think of a siciliano is as a slow gigueA telltale sign of a siciliano is often a large number of dotted rhythms. [17]  An example of this can be seen in Michelangelo Abbado’s edition of the Devil’s Trill Sonata as shown below.  (Figure 3)[18]


Figure 3 Andante from Abbado's Edition of the Devil's Trill Sonata showing the dotted rhythms typical of a Siciliano as well as the time signature.


One of the most drastic changes that took place in Kreisler’s edition of the Devil’s Trill Sonata was the elimination of most of the double stops in the opening Andante section.  (Or what Kreisler referred to as the Larghetto section.)  (Figure 4) [19]  Joseph Szigeti comments upon the loss of the double stops in the first movement of the sonata in his book Szigeti on the Violin

“Even Tartini’s important and perennially popular ‘Devil’s Trill’ Sonata has not escaped editorial distortion.  Of all the editions at present available only two have preserved the original double-stopped statement of the lovely Siciliana-like first movement…we take for granted the suppression of the necessary and enhancing accompanying lower voice.”[20]



Figure 4 The first two measures of the Kreisler edition of the Devil's Trill Sonata.  Notice how all of the double stops have been omitted.


       Kreisler’s omission of the double stops in the first movement of his edition may have happened for several reasons.  The harpsichord, or the cello, that would have been used to accompany Tartini may not have been able to perform these notes and bring them out as Tartini felt was necessary or Tartini felt that the violin should accompany itself since it was the quiet beginning of a siciliano.  Tartini, in my opinion, may have used the double stops throughout the siciliano as a way of adding to mysterious and yet pastoral quality of the sonata.  More importantly, Tartini’s violin and bow were accustomed to playing double stops such as these.  In fact, the violin bridge and bow of the baroque era were adept to rolling chords and making them sound in ways that are very difficult and awkward to do on the modern violin.  By looking at figure two, it is obvious that the bridge on the baroque violin was less rounded than its modern counterpart, which meant that there was a lesser degree in difficulty in playing double stops when compared to the modern violin.  It can be easily understood why Kreisler omitted the double stops, and used the piano accompaniment part to fill in the chords that he had removed from the violin part.

Refraining from the use of double stops was not the only change that Kreisler made to his edition of the Devil’s Trill Sonata.  Period performance practices greatly influence many of the changes that Kreisler ended up making to his edition.  One very striking example of this is in his use of placing certain passages or phrases completely on one string to add to their character.  This was a common practice in place during the Romantic era and now even into the twenty-first century. 

During Tartini’s time, many violinists would have been able to play such passages on a single string.  However, the drawback to playing such a passage on a single string would have meant the sound quality was sacrificed because of the shortened length of the string.  Baroque violinists much preferred the sound of a longer string in their playing, which meant that they would have stayed in the lower positions as much as possible in order to retain a better sound quality.  A good example of the Romantic idea of fingering a passage or phrase on the same string is shown below (Kreisler’s version) and is compared to what would have been acceptable performance practice during the baroque era (Abbado’s version).  (Figures 5 and 6) [21]


Figure 5 Kreisler's version of the Devil's Trill Sonata, measures 5-10 (measures 4-12 shown).  Notice the fingerings that are printed in, starting with the second measure on the first line



Figure 6 Abbado's version of Tartini’s Devil’s Trill Sonata, measures 5-10. Notice the difference in fingerings present from the above example.  Also, pay close attention to the similarity between the marked in bowings on the first three notes of m.5 when compared with the same place on figure 5.



Bowings in figures five and six above provide an interesting contrast when compared each other, even though the same bowing is being shown.  As was mentioned earlier, the baroque bows did not produce an even tone from the frog to the tip.  It is because of this fact that Kreisler’s bowing is so interesting.  The violin bow that Kreisler would have had in mind while writing his version of the sonata would have created a very even tone when drawn from frog to tip.  Kreisler’s intentions were good when he left the bowing in m.5-10 very much like what had been in the manuscript version of the sonata, however, because the bows created such different sounds, Kreisler ended with a very smooth and sweet sound.  While Abbado’s phrasing and marked bowings appear to create the same effect, they do not because of the manner in which the baroque bow creates sound.  A consonant sound was produced when the bow was first pulled across the string, on the up bow a vowel sound was created and again on the down bow a consonant sound was created.  The first and the last notes of the three note phrase in m.5 received more emphasis than the middle note simply because of the difference between the baroque and modern bows.  This particular phrase finds itself repeated throughout the sonata with similar treatment as mentioned above by Abbado and Kreisler respectively.

In the two musical examples shown below, a comparison between notes yields some interesting comparisons and contrasts.  (Figures 7 and 8)[22]   What proves most interesting is in the fact that in figure 8, Abbado’s example, contains two markings with stars (one with one star, and another with two stars).  Both of these markings are used to show that those particular passages are exactly what are in the manuscripts of Tartini’s students.  When Kreisler’s version is compared with that of Abbado’s version, especially at these two places, it can be seen that he changed the notes, the rhythm, and bowings.  Most likely, Kreisler did not have the original manuscript from Tartini’s students, but rather was using music that had been edited by one of the great nineteenth century virtuoso violinists such as Joseph Joachim (1831-1907) or Henri Vieuxtemps (1820-1881).[23] 

There seem to be three places where there are drastic differences in notation between Abbado’s and Kreisler’s versions of the sonata.  The first of these is m.21.  Kreisler has changed the rhythm of the three double stop chord at the end of this measure, from one being an eighth note and a dotted quarter note, to being two quarter notes and an eighth note.  By doing this, Kreisler has put greater emphasis on the last two and a half beats of the measure than Tartini had.  In Tartini’s version, a breath before the last two beats of the measure is used to emphasize these chords.  Kreisler used the extra quarter note chord just as Tartini used the breath before the two chords at the end of the measure to emphasize these chords in a very Romantic manner. 

Two of the most interesting changes that Kreisler made to his version of the sonata come from the manuscript of the Devil’s Trill Sonata in the hand of Tartini’s students.  Kreisler uses the same technique to emphasize the chords in m.22, and in doing changes the rhythm once again.  In doing this, Kreisler wanders away from the manuscript version of the sonata completely and even goes so far as to double the emphasis at the beginning of m.22 by making a chord of the note d.  This is in itself ironic because Kreisler eliminated most of the double stops within this section of the sonata, yet here he is adding one.  In m.23, Kreisler repeats the same figure at the beginning of the measure as he did in measure twenty-two.  Kreisler then goes on to change the rhythm at the end of the measure, but not without first eliminating the double stop used to end this section of the sonata.

Along with drastic differences between these three measures, there is also an interesting concept present in the bowing.  Like the consonant and vowel sounds that were discussed earlier, Kreisler found a way to emphasize the notes in m.21 in a similar manner as the bowing does at m.21 in Abbado’s version of the sonata. 

Figure 7  Kreisler’s Version of the Devil's Trill Sonata, m.21-23 (measures 19-23 shown).  Pay close attention to the subtle differences between the rhythm and notation here and then compare with figure 8. 



Figure 8  Abbado's Version of the Devil's Trill Sonata, measures 21-23 (measures 19-23 shown).


            During the lifetimes of Joachim, Kreisler, and Vieuxtemps, the modern violin bow was considered an improvement upon the baroque bow (and the baroque bow was considered obsolete), rather than being considered a different bow meant for different purposes in violin playing.  The modern violin bows that we have today cannot replicate the sounds or such effects as the messa di voce as well as their baroque counterparts.  Of course, we can attempt to replicate the sounds, but we are asking both the instruments and bows that are being used to do something for which they were not made. 

          Even the bowings in the figures shown above have been changed in manner that reflects the differences in bows.  In addition, these differences between bowings change the entire emphasis and character of these passages.  Abbado keeps the bowings smoother in general, while Kreisler prefers to have more of an attack upon the notes by using staccato and heavily accented bowings.  Another example of the differences between the bowings of the two versions is shown below.  (Figures 9 and 10)[24]  Once again, it seems as though Kreisler takes the offensive with the bowing as to present more of an attack on the notes in contrast with Abbado’s version of the sonata.  (There is a mistake in measure thirty-nine of the sonata, where Kreisler forgot to place a flat on the high note e since he had an accidental on the low note e in the beat before.)


Figure 9 Kreisler’s version of Tartini’s Devil’s Trill Sonata, measures 39-48 (measures 37-53 shown).  Notice the differences in bowing between this example and the example below.



Figure 10 Abbado’s version of Tartini’s Devil’s Trill Sonata, measures 40-49.  Observe the differences between the bowings shown above and those shown here.  Many of the sixteenth notes are in pattern of being slurred as groups of three or two, in comparison with above where slurs are only used with groupings of two notes.



            Both of the above examples are typical of the differences in this type of sixteenth note passage throughout the sonata.  In Kreisler’s version there does not seem to be a difference between the idea of the trill and the idea of the nachschlag.  A nachschlag is in reference to the suffix of a trill, meaning that the note above the marked note would be played within the written notation.[25]  This difference within the markings of these two versions of the sonata brings up the difficult question of ornamentation.  Ornamentation during the baroque era was something that was considered a way of a performer putting a personal fingerprint on a performance that they were giving.  Not everyone could write his or her own ornaments for the music being written during Tartini’s lifetime, however if someone who was not willing to write ornaments could have found a copy of a sonata with ornaments already added. 

          The first known publication of the Devil’s Trill Sonata was in 1798 as part of Cartier’s method, L’Art du Violon ou Collection Choisie dans les Sonatas des Écoles Italienne, Française et Allemande. According to Ginsburg, this sonata had fallen into obscurity until 1855 when Henri Vieuxtemps and Robert Volkman published an arrangement of the sonata.[26]  By the time this eighteenth-century sonata had become popular again it was the Romantic era and performance practices dictated that the notation on a page of music was a parameter to be followed, there was no room for ornamentation.  What is known about the ornamentation that Tartini may have used is found in his Traité des agrémens de la musique.  It is here that Tartini discusses ornamentation and refers to the analogy of a cook adding too much or too little salt (with the salt being the ornamentation).[27]

            The figures above (9 and 10) showing Abbado’s use of the nachslag and Kreisler’s use of the trill give the perfect example of the very different times in which these two versions of the Devil’s Trill Sonata were written.  Subtle differences between the use of these two ornaments here, in addition to many other sections within the sonata open up a lot of questions that have to do with the personal preference of the performer and are not appropriate to deal with in the same manner as issues with bowing, fingering, and changes to notation.  It is very likely that the same performer never used the same set of ornaments twice in a row.  There is no doubt that those ornaments would have been in the same place in the music, but even the idea of where to ornament changed throughout the Baroque era. 

          Like ornamentation, the idea of the cadenza for this sonata is of personal preference.  The manuscript of the Devil’s Trill Sonata attributed to Tartini does not include a cadenza.  Both Abbado’s and Kreisler’s versions of the sonata do contain cadenzas with the note that they were written by the arranger/editor.  The best idea for performance practice, whether it be performance practices for the Baroque era or the Romantic era, would be for the performer to write his/her own cadenza as they see fit.  Cadenzas were meant for the musician to be able to show the audience the techniques that they were exceptionally good at on an instrument.  The only kind of cadenza that can the goal of the above-mentioned cadenza, as it is discussed here, is one written by the performer and played by the performer. 

            Understanding the differences between the performance practices of the baroque era when compared to those of later times is crucial for any musician.  Repeatedly, students and professors alike who have studied historical performance practices have remarked on how it has influenced how they view and treat music in general.  Kreisler’s version of the Devil’s Trill Sonata has a place in history just as much as the manuscript version of the sonata that is attributed to Tartini (but not in Tartini’s hand).  In order to understand the beauty of the manuscript version, it is necessary to delve into such things as the violin methods and techniques of the past, the development of the instrument, and the background present to the creation of the work.  David Boyden comments upon this very idea in his article entitled “The Violin and its Techniques in the 18th Century.”

“The importance of the information in the violin methods consists simply in this:  these old books reveal the kind of beauties that lie concealed under the plain notes of the scores, particularly with respect to matters of bowing, expression, ornaments (including vibrato), and dynamics.”[28]


          The above quote is true for all music composed for the violin, whether it was composed merely a hundred years ago or closer to two hundred and fifty years ago.  Kreisler’s edition provides a glimpse into the world of Viennese charm and warmth that seems to be ever present in many of his works.  This in itself is a great historical influence that should be considered just as powerful as the circumstances under which Tartini is said to have written his version.  Tartini, Kreisler, and many others who have edited/arranged versions of this sonata have left a historical fingerprint by what they have done.  Let them all be considered in equal light when placed against the performance practices of their time, because each version, including the manuscript version attributed to Tartini, is a lesson in the history and development of the violin and its music.



Copyright © 2001 Michelle Edelman. All rights reserved.

To Top

[1] Lev Ginsburg, Tartini His Life and Times, (Neptune City, NJ:  Paganiniana Publications, 1981), 117.


[2] Joseph Jerome de Lalande, Voyage d’un Francois en Italie, fait dans les Annes 1765 et 1766.  Tome huitme, (Paris:  1760), 293-294.  As quoted in Lev Ginsburg’s Tartini His Life and Times, 111.

[3] Lev Ginsburg, Tartini His Life and Times, 187.

[4] Lev Ginsburg, Tartini His Life and Times, 111.

[5] Louis Lochner, Fritz Kreisler, (New York:  The Macmillan Company, 1950), 17. 

[6] Joseph Szigeti, Szigeti on the Violin, (New York:  Dover Publications, Inc., 1979), 140.

[7] Henry Roth, Violin Virtuosos from Paganini to the Twenty-first Century, (Los Angeles:  California Classic Books, 1997), 41. 

[8]Robin Stowell, The Cambridge Companion to the Violin, (Cambridge:  Cambridge University Press, 1992), 24.

[9] Robin Stowell, Violin Technique and Performance Practice, 166.

[10] David Boyden, The History of Violin Playing from its Origins to 1761, (London:  Oxford University Press, 1965), 393.

[11] Robin Stowell, Violin Technique and Performance Practice, 166.

[12] F. J. Fétis, Antoine Stradavari, luthier célèbre, (Paris: 1856), 74.  As shown in Robin Stowell’s Violin Performance and Practice, 15.

[13] Robin Stowell, Violin Technique and Performance Practice, 23.

[14] Robin Stowell, Violin Technique and Performance Practice, 25.

[15] Ibid, 23.

[16] The New Harvard Dictionary of Music (1996), s.v.  “Sonata da Chiesa,” by Stephen Bonta.

[17] The New Harvard Dictionary of Music, s.v.  “Siciliano,” by Bruce Gustafson.

[18] Michelangelo Abbado, Sonata in sol minore “Il Trillo del diavolo,” (Milano:  Ricordi, 1974), 1.

[19] Fritz Kreisler, Sonata in G minor “Devil’s Trill,”  (New York:  International Music Company), 1. 

[20] Joseph Szigeti, Szigeti on the Violin, 138.

[21] Fritz Kreisler, Sonata in G minor “Devil’s Trill,” 1; Michelangelo Abbado, Sonata in sol minore “Il Trillo del diavolo,” 1.


[22] Fritz Kreisler, Sonata in G minor “Devil’s Trill,” 1; Michelangelo Abbado, Sonata in sol minore “Il Trillo del diavolo,” 2.

[23] Lev Ginsburg, Tartini His Life and Times, 117.

[24] Fritz Kreisler, Sonata in G minor “Devil’s Trill,” 2; Michelangelo Abbado, Sonata in sol minore “Il Trillo del diavolo,” 3.


[25] The New Harvard Dictionary of Music, s.v.  “Nachschlag,” by David Fuller. 

[26] Lev Ginsburg, Tartini His Life and Times, 117.

[27] Giuseppe Tartini, Traité des agrémens de la musique, (Celle & New York, 1961), 74.  As quoted in Robin Stowell’s Violin Technique and Performance Practice, 322.


[28] David Boyden, “The Violin and its Technique in the 18th Century,” The Musical Quarterly, January 1950.