J. S. Bach’s Prelude and Fugue in E Flat (BWV 552,1/2): An Inspiration of the Heart?

[1] After many years of performing Clavier-Übung III—Bach’s catechism cycle of chorales which is framed by the opening Prelude in E♭ and the concluding Fugue in the same key—my interest in the deeper underlying issues embodied in the collection was rekindled upon reading Gregory Butler’s penetrating analysis of its genesis.⁽¹⁾ His proof that Bach’s intentions underwent a dramatic change at the engraving stage (which caused a serious delay in the production and publication), and his speculation as to some probable musical influences from Hurlebusch which can be detected in the Prelude and Fugue movements, led me to ask one simple question: why? Why did Bach change his mind about the work’s conception? And why in effecting this change did he wish to insert music based essentially upon the ideas of another composer?

[2] Despite the late changes, the title page was left unaltered. Furthermore, Butler has demonstrated not only that it must have been engraved early on, but also that it must have been the very first plate engraved for the collection.⁽²⁾ Since this page was left unchanged, whether through expediency or other reason, we must at least thereby assume that whatever the motive for changing his mind may have been, Bach’s eventual purpose was in essence no different from what it had originally been: he intended, as the title states, to present “various preludes on the catechism and other hymns for the organ.” It is therefore with Luther’s teaching that I wish to begin a further investigation of Bach’s motives as displayed in the final version.

[3] The word “catechism” means “instruction manual,” and the usual manner of presentation is in question-and-answer format. Of Luther’s two Catechisms, only the Shorter follows this structure; the Larger is really a series of sermons and expositions. Luther describes the word twice:⁽³⁾ a) “a compend and brief summary of all the scriptures” (Preface to the Larger Catechism); and b) “instruction for children” (Preface to the Shorter Catechism). Bach’s collection of catechism pieces could therefore be viewed on two concurrent levels: a) as a collection of musical items for “ad libitum” use in Luther’s liturgy; and b) as a didactic cycle reflecting salient aspects of Luther’s own meditations upon this. Thus, not only has Bach provided interesting chorale settings that bear upon the liturgy which Luther outlined, and with which they can be used “ad libitum” to enhance its realization, but also his structural plan has reflected—in its provision of longer and shorter settings—the two respective catechisms themselves.

[4] A catechism is, however, entirely instructional: it does no more than provide an explanation of, and authority for, the believer’s faith and its necessary means of expression. In Luther’s catechism, moreover, there is an extra dimension of importance: a dissatisfaction with purely orthodox tenets, and an affirmation of a reformed mode of thought. This places as much emphasis upon the spirit as upon the letter of the law.

[5] The possibility that Bach’s agenda in Clavier-Übung III was to embody in his musical expression not only the letter but also the spirit of Luther’s writings makes it worth examining further those movements which were added at a later stage. In particular the opening Prelude and concluding Fugue, neither of which bears any clear external relationship with Luther’s teaching, are prime candidates for review. I hope to demonstrate that both movements, despite their format, are linked both inspirationally and musically to a common chorale source; and that this chorale itself was reflective of a tangible Lutheran premise which had acquired a prominent position in Luther’s Larger Catechism.

[6] Such a proposal will need to establish the three attributes of all serious litigation: means, motive and opportunity. Should all three be deemed satisfactory, it will then further be necessary to provide explicit and corroborative evidence—via analytic means—in order to arrive at a successful verdict. This essay will therefore proceed under the following headings: Means, Motive, Opportunity, Evidence.

[7] MEANS: Although the basing of a Prelude and Fugue on a chorale may seem fanciful, there is at least the evidence that fughettas utilized this procedure. A particularly salient example is the fughetta on Allein Gott (BWV 677) from Clavier Übung III. Examination reveals that the designated chorale lies buried within the subject and its extension, and that the second phrase of the melody is treated in a similar manner (see Example 1).

[8] The particular pertinence of this example lies not so much in its origin in Clavier Übung III as in its close musical association with the Fugue in C Major (BWV 547) whose opening can scarcely be viewed as anything less than a conscious elaboration of the same material (see Example 2).

[9] With this parallel, it is difficult to believe that Bach’s association of the chorale with the fughetta would not have been matched by a similar connection with the fugue. The former is a clear designation, while the latter is an elaborate expansion of the very same material. Looking now at the Prelude in C—the fugue’s companion—is there further evidence of a material connection with the chorale? Example 3 may afford some insight:

[10] These affinities between the Prelude and the chorale are maintained throughout the movement, although the association is rhapsodic rather than schematic. The Prelude here develops its own free-standing motivic content; but the origin of these motifs bears more than a passing reference to the chorale. The means to incorporate chorale-derived material into Preludes and Fugues seems, on the basis of Examples 2 and 3, credible as opposed to merely fanciful.

[11] MOTIVE: Certain aspects of Luther’s catechism are innovatory and significant. In particular may be mentioned his teaching of the Creed in Part 2 of the Larger Catechism where he rationalized the twelve Articles of the Creed into three “according to the three persons of the Trinity”: God the Father (creation), God the Son (redemption), and God the Holy Ghost (sanctification).⁽⁴⁾ This is not merely a structural simplification of pre-existing dogma, but a theological rationalization of Belief and its essential attributes. Bach’s desire to express this in the structure and design of the cycle, and especially within its concluding fugue, also indicates his interest in the doctrinal foundations and theological concepts which Luther’s instruction manuals sought to illuminate.

[12] But Luther’s discussion of the Creed was only subsequent to his exposition of the Ten Commandments; it was his view that before we could genuinely state our beliefs, we had to understand the Law: “Thus far we have heard the first part of Christian doctrine (i.e. the Ten Commandments), in which we have seen all that God wishes us to do or to leave undone. Now, there properly follows the Creed, which sets forth to us everything that we must expect and receive from God. . .”

[13] In his conclusion to the Creed, Luther sought to contrast its nature with that of the Ten Commandments: “. . .the Ten Commandments are written in the hearts of all men; the Creed, however, no human wisdom can comprehend, but it must be taught by the Holy Ghost alone.” This allusion to “the heart” is of great importance to Luther’s view of the Commandments, and man’s adherence to them. Indeed, his conclusion to Part I (Ten Commandments) states the following: “Thus He demands that all our works proceed from a heart which fears and regards God alone. . .” Luther saw “the heart” as being axiomatic to his new philosophy of our interaction with God, for in Article III on The Lord’s Prayer he stated: “And if we would assemble all the churches, together with all ecclesiastics, they would be obliged to confess that they never from the heart prayed for even a drop of wine.”

[14] References to “the heart” in Luther’s commentary on the Ten Commandments (Larger Catechism Part I) are too frequent to enumerate here without tedium; but a few selected examples will, hopefully, serve to illustrate this crucial attribute of correct observance required by Luther:

i) “. . .to have a God is nothing else that to trust and believe him from the whole heart; as I have often said that the confidence and faith of the heart alone make both God and an idol.” (First Commandment)

ii) “As the First Commandment has instructed the heart and taught faith, so this commandment leads us forth and directs the mouth and tongue of God.” (Second Commandment)

iii) (Speaking of the training of children) “But this manner of training so spreads its roots in the heart that they fear God more than rods and clubs.” (Conclusion to the Second Commandment)

iv) “. . .but, most of all that both in heart and with body we so act as to show that we esteem them (i.e. the parents) very highly.” (Fourth Commandment)

v) “When we see such people (i.e. our enemies), our hearts, in turn, could rage and bleed and take vengeance.” (Fifth Commandment)

vi) “. . .and not only is the external act (i.e. of adultery) forbidden, but also every kind of cause, incitement, and means, so that the heart, the lips and the whole body may be chaste. . .” (Sixth Commandment)

[15] Thus the Creed, where we acknowledge our belief in the gifts of the three persons of the Trinity, is only a corollary to an adherence to the Ten Commandments which are written (not just on stone tablets but) in our hearts: while prayers and beliefs are spoken with the lips, they must originate from the heart.

[16] OPPORTUNITY: If Bach had wished to capture Luther’s emphasis on “the heart,” he would probably have done so by allusion. An unambiguous homage to this emphasis could, for example, be paid by presenting a composition—integral to the overall scheme—which has strong and clear affinities with “the heart” in the minds and ears of those studying or hearing the piece. Such an opportunity would be presented by a chorale, well-known to congregations of the time, which focussed in some way upon this very aspect. If they recognized the use of the chorale melody, either through musical quotation or borrowing, they would also call to mind the appropriate text associated with this melody.

[17] One clear opportunity, with appropriate associations, is the chorale O Herzens Angst, O Bangigkeit.⁽⁵⁾ The fact that the version which has come down to us, with Bach’s own harmonization, uses the same key as the Fugue is less important than two other features: a) its text⁽⁶⁾ presents a clear link with the heart, and b) its first four notes are identical with those of the main fugue subject. Whether this instant but initially superficial connection will lead to a more tangible inbuilt association between the two must now be discovered by examining, through analytic means, the evidence.

[18] EVIDENCE: The fughetta on Allein Gott earlier showed how Bach, in assimilating a chorale melody into a fugal texture, embroidered its main notes into a highly-decorated new theme. It also showed that, unlike in the more formal, designated, chorale settings, the treatment of the model is freer in nature so as to allow the motivic content of the new piece per se to develop along its own lines as is customary in fugues (whether on a small or large scale). But the very title of the fughetta compels an acceptance of the presence of the melody, and a simple analysis of the structure reveals how this is used. The main difference between this fughetta and the larger Fugue in C (BWV 547) is, then, that our perception of this relationship in the former is initially and primarily external (i.e. it is designated in the title), and that we are then inexorably led to explore how this relationship has been internalized through the music. With the Fugue in C there is no external connection; the clear internal relationship with the fughetta does, however, strongly impute a corresponding internal relationship with the same chorale.⁽⁷⁾ Any evidence pertaining to the Fugue in E Flat must, therefore, be purely internal: the starting point does not arise from external evidence provided by the composer, but rather from the circumstantial evidence deduced earlier with regard to means, motive and opportunity.

[19] The fugue subject itself is not one of striking innovation: a very similar one can be found in the organ works of Buxtehude,⁽⁸⁾ and a case was recently argued in favour of a direct link with Hurlebusch’s Compositioni Musicali (specifically the fugue number 4 of his Parte Seconda).⁽⁹⁾ While it is important to reveal such likenesses, it is no less crucial to demonstrate other impulses which seem to have—as I would argue in this instance—a clear conceptual relevance to the whole cycle (of which this fugue is but one important member).⁽¹⁰⁾ In this respect, I hope to show that Bach’s example contained far more musical, theological, and spiritual resonance than all the other examples put together.⁽¹¹⁾

[20] The text of O Herzens Angst strikes a number of resonances with Luther’s Larger Catechism, not least with regard to admonitions contained in his commentary on the First Commandment.⁽¹²⁾ Those unequivocal Lutheran axioms evidence just how strongly a Lutheran heart must experience both love and fear. This chorale is illustrated in Example 4, together with the surviving text:

[21] Upon examination, the opening of the E-Flat Fugue displays a clear similarity with the beginning of the chorale; what may not be so evident, however, is that the countersubject of the fugue also shows a strong resemblance to the second phrase of the chorale. But in both themes there are extra notes: the subject is provided with a cadence figure in measure 2,⁽¹³⁾ while the countersubject provides added pitches at the end of its first measure in order to retain movement and harmonic integrity with the answer.⁽¹⁴⁾ The various correlations between the fugue’s beginning and the chorale melody are illustrated in Example 5:

[22] The opening of the fugue’s second section (measure 37) presents what initially appears to be a new subject: but closer examination reveals that it is nothing other than an elaborate variant of the original countersubject (see Example 5, measures 3–4, relating with the chorale’s second phrase). So far, therefore, it can be seen (even supposing this to have been purely subconscious on Bach’s part) that the fugue employs, in its first section, material closely related with the first two phrases of the chorale. If, however, one were to discover, as the fugue progresses, further references to the chorale melody articulated in a definite and schematic manner, this would begin to suggest that the composer was not operating only on a subconscious level but consciously deriving essential structural features from the chorale. Example 6 shows the opening of section 2, and the accompanying line supplied from the chorale’s second, third and fourth phrases:

[23] That this new subject should correlate with phrase 2 of the chorale is fairly inevitable since it is a variant of the first countersubject (which also follows the contours and main pitches of the chorale); but there can surely be nothing coincidental in its continuation not only ghosting phrase three but also phrase four of the chorale. As we know very well, measure 59 will herald the return of the first fugue subject at its original pitch (to be combined then with the new subject), and this will thereby recall phrase 1 of the chorale (together with its cadence figure which ghosts the final chorale phrase).⁽¹⁵⁾ So far, therefore, section two of the fugue has shown a relationship with phrases two, three and four of the chorale melody, and this has followed a simple presentational scheme whereby within only five measures almost half the entire model has been quoted. When the fugue’s original main theme returns at measure 59, this section will then have quoted from the first four of the chorale’s six phrases.

[24] As the second section of the fugue moves to its close (by measure 82), its tonality begins to veer in the same direction as phrase five of the chorale: it moves from B Flat Major (measure 59, pinpointed in structure by the first return of the original theme) to C Minor (measure 82) where a firm cadence is established. Since, however, it seems that Bach has not quoted directly from the melody of phrase five,⁽¹⁶⁾ any supposed allusions will not strictly satisfy the forensic criteria stipulated in this essay. Measures 75 to 82 are packed with sophistication involving the returned original main theme, and variants of the new main theme in both rectus and inversus positions. In this pregnant outpouring of technical skill, in which the music is carefully pushed towards the C Minor goal, it is nonetheless possible to hear a number of musical references, even though they can at best be regarded as only “circumstantial” in nature. These are illustrated in Example 7.

[25] At measure 82, the fugue moves effortlessly and without a break into its third and final section. A new, third subject emerges from within the cadence, beginning with the notes C–B♭–C. These are also the same notes that are found in the chorale as it moves from its C Minor cadence into the sixth and final phrase.⁽¹⁷⁾

[26] In conception, the external plan of the fugue is transparent; but I have argued that there is also an internal plan which links it to a particular chorale. This connection arises not because of the visually obvious fact that each composition begins with the same four notes; viewed merely on this level, the true plan remains invisible. Only by making an important connection with Bach’s Lutheran outlook—one that intersects also with the text of this particular chorale—is one led beyond these four notes to perceive the extent to which these two pieces musically interact. This connection—the imagery of “the heart”—may possibly suggest a reason as to why Bach remained cautious and tardy about his final, published conception of the work, even well into its period of engraving. Perhaps he considered at the last moment that an important spiritual aspect of Luther’s teaching had not at that stage found an adequate formal means of expression within the scheme proposed.

[27] As Gregory Butler has conclusively proved,⁽¹⁸⁾ the opening Prelude in E Flat was added after the engraving of the collection had been begun; but why? Was there also an internal agenda for this movement? And if so, did it follow that which appears in the fugue? Indeed, is there any direct musical connection between the two movements (over and above the purely superficial one of key)?

[28] Butler has argued in favour of a link between the Prelude in E Flat and the Overture of the Parte Seconda in Hurlebusch’s Compositioni Musicali.⁽¹⁹⁾ While, again, we see Bach’s monumental musical conception set against Hurlebusch’s more modest accomplishment, a similarity of style cannot be denied; but these purely external attributes, again, do not necessarily explain what has been internalized or implanted beneath the surface detail. While no definite musical link has yet been posited for the two Hurlebusch movements (other than their being separate items within the same collection), I would argue that there certainly is internally a unifying musical (and philosophical) link between Bach’s two movements.

[29] Most modern editions of the Prelude in E Flat transmit the music on three staves for ease of performance. This presentation obscures what I believe is an important structural element in the opening 7 measures: an element that becomes transparent when the original print, which uses only two staves, is viewed.⁽²⁰⁾ There, an important melody in the left hand is laid bare for all to see, as shown in Example 8.

[30] This melody has at least a clear visual connection with the first theme of the concluding fugue. Aurally, it is less obvious because a) the rhythmic scheme of the opening requires note repetition, and b) the ear is directed more to the Prelude’s main theme in the right hand. Other connections can be perceived, albeit with only varying degrees of forensic corroboration, between the Prelude and the chorale melody.

[31] Like the chorale, the opening ritornello follows a simple harmonic plan: E♭–B♭–C minor–E♭. In the chorale, the move from B♭ to C minor is effected melodically by the distinctive rising chromaticism of B♭–B♮–C. In the Prelude (measures 17–20) the same line is found in the bass.

[32] More importantly, however, the second ritornello (measures 51–71) reaches its final cadence in C minor, and the continuing “solo” material now places an emphasis upon what appears to be a new theme: one that recurs with considerable development later in the piece. This new theme has very close links with phrase 4 of the chorale: it occurs in both principal contrapuntal lines. After extensive development through diverse keys, it eventually returns in C minor (measure 169) to close the final “solo” section. Measures 71–5 are shown in Example 9.

[33] CONCLUSION: Bach was hesitant about the final form of Clavier Übung III: this was either because he was dissatisfied with the quality of some of the music he had written, or because he became skeptical about the presentation and format as originally conceived. Since Butler has proved that the changes largely involved the addition rather than the removal of material, the latter reason seems the more likely. The only explanation for wishing to enlarge the collection—which decision caused enormous problems with the scheduled production and publication dates—must be the way in which Bach viewed his success in having fulfilled the overall objective: one of providing musical items reflecting Luther’s Catechism. It is easy for such a scheme to follow the letter of the law (i.e. through the provision of settings of the Ten Commandments, the Creed, and the Sacraments), but not so easy for it to capture the spirit. Luther’s teaching stands apart by virtue of its clear insistence that The Law must embody a human experience “from the heart.” All such experience, whether painful or pleasurable, needs this essential attribute.

[34] The opportunity for Bach to capture this nuance arose, I would argue, through the existence of a chorale whose text lent an appropriate emphasis for a composer able to seize upon the musical elements associated with this.

[35] The means of utilizing such musical elements within the framework of a Prelude and Fugue already existed through the technique currently used in embedding chorale melodies within fugal textures (e.g. in the designated chorale-fughuettas and fugues), and in the establishment of thematic links between Prelude and Fugue.

[36] The motive presents itself in Bach’s wish (after a period of re-think and procrastination) to capture not only the letter but also the spirit of Luther’s teaching as transmitted through his catechisms. Luther required observance of the law to be “from the heart”; and I have argued that this special and significant attribute was conveyed by Bach in opening and closing movements which, through their clear spiritual resonances, provided a fitting envelope for a collection of music inspired by Luther’s catechism.

Roger Wibberley
Goldsmiths University of London
Music Department
New Cross
London SE14 6NW
mua01rw@gold.ac.uk

Footnotes

1. Gregory Butler, Bach’s Clavier-Übung III: The Making of a Print (Durham and London: Duke University Press 1990).
Return to text

2. Butler, 55.
Return to text

3. The translations from Luther used throughout this essay are taken from F. Bente and W. H. T. Dau, Triglot Concordia: The Symbolic Books of the Ev. Lutheran Church (St. Louis: Concordia Publishing House 1921). These texts are in the public domain via the internet at the following URL: http://www.iclnet.org/pub/resources/text/wittenberg/wittenberg-luther#lc
Return to text

4. Luther concluded: “One God and one faith, but three persons, therefore also three articles or confessions.” The trinitarian aspects of Clavier-Übung III (seen in the number of movements, the tripartite design of the fugue, its number of entries of the subject, and its signature of three flats) seem to mirror Luther’s view that belief, and its expression in the Creed, is in essence symbolic of, and derivative from, the Trinity.
Return to text

5. See Albert Riemenschneider, 371 Harmonized Chorales and 69 Chorale Melodies (New York/London: Schirmer 1941), no. 173. No J. S. Bach source survives for this chorale, though it appeared in C.P.E. Bach’s edition of the chorales, vol. 2, published by Breitkopf in 1785.
Return to text

6. See Carl von Winterfeld, Der Evangelische Kirkengesang III (Leipzig: 1847; reprint ed., Hildesheim: Olms 1966, 287, where the text is attributed to Christi Begraebniss). No texts were transmitted by C.P.E. Bach.
Return to text

7. It would be difficult to maintain that Bach, even if Allein Gott could not have been further from his mind when he started work on the Fugue in C, did not at least become aware of the musical connotation as his work proceeded, especially as the material he used was almost identical to another piece where he provided a clear external link. The further evidence internal to the accompanying Prelude in C can only serve to strengthen this assumption.
Return to text

8. See the Prelude and Fugue in E Major in Hedar, Josef, Dietrich Buxtehude, Saemtliche Orgelwerke, Band II (Copenhagen: Hansen 1952), 80.
Return to text

9. Butler, 5.
Return to text

10. Butler’s study of certain similarities between the Hurlebusch and Bach compositions (including important historical facts providing evidence that Bach was acquainted with his music) is interesting and informative. However, Hurlebusch’s piece, despite its alla breve notation, seems stylistically more akin with the E-Major fugue of Buxtehude. The similarities extend beyond the subject (the second half of which is almost identical in both cases, though radically different in the Bach): although Hurlebusch at least surpassed Buxtehude in realizing the possibility of simple stretto (Buxtehude merely hinting at the possibility once in measures 31–2), the structure of his exposition is somewhat loose compared with Bach’s. In Hurlebusch, not only does the second voice announce a statement of the theme again in the tonic (rather than the dominant), but the composer cannot withhold the temptation of using stretto as early as measure 7. Bach leaves it until measure 21 where the device marks off a structural point in the first section of his fugue. This contrast alone, while not negating the proposed link with Hurlebusch, does point to fundamental differences in the respective concepts of overall structure and scale. Further, Bach (unlike either Hurlebusch or Buxtehude) alters the pitch of note 4 in the answer; this idea was not influenced by either of the other composers, but stems from the salient importance of his countersubject (which I will demonstrate presently).
Return to text

11. Nothing in this essay seeks to challenge Butler’s findings with regard to the genesis of Claver-Übung III. It may, however, provide a possible reason as to why Bach changed his view of the work’s overall design even after engraving had begun.
Return to text

12. The chorale text may be translated as follows: “O fear of the heart, O anxiety and hesitation! Whose body do I see being carried? Whose tomb is this? What is the name of the rock? I should know it!” Luther made it clear in his teaching on the First Commandment that man should both love and fear God. He wrote in the Exposition of the Appendix to the First Commandment: “Learn, therefore, from these words how angry God is with those who trust in anything but Him, and again, how good and gracious He is to those who trust and believe in Him alone with the whole heart. . . He is a God who will not leave it unavenged if men turn from Him. . .even until they are utterly exterminated. Therefore He is to be feared, and not to be desisted.”
Return to text

13. This matches—interestingly—the final cadence of the chorale melody.
Return to text

14. The countersubject is (despite appearances) not a regular one: its first statement is crucial to the unfolding of material, so much so that Bach is happy to alter the pitch of note 4 in the answer. This alteration also occurs in measure 10 when this countersubject (with one minor change of pitch) reappears.
Return to text

15. This will be the first time when the cumulative structure of the whole fugue can be perceived: the old theme returning to combine with the new also happens in the third section. This will identify the original subject as, indeed, the overriding “main” theme. Its association with the words “O Herzens Angst” will therefore be an ever-recurring one throughout the piece.
Return to text

16. This is because his idea for this section is to show the new subject both in its rectus and inversus forms. It was argued above that even where fughettas bear chorale designations in the title this does not compromise the music’s primarily fugal design.
Return to text

17. These notes, in Bach’s own harmonization, are also echoed in the tenor voice. Following these three notes, the chorale contains only a single melodic passing note before reaching the E♭ note of the final cadence; and this was the fragment which Bach chose at the outset to complete his original subject. It therefore returns along with the main theme’s reappearance to round off the whole movement.
Return to text

18. Butler, 39–43.
Return to text

19. Butler, 4.
Return to text

20. For a facsimile, see NBA vol. IV/4, xi.
Return to text

Copyright Statement

Copyright © 1998 by the Society for Music Theory. All rights reserved.

[1] Copyrights for individual items published in Music Theory Online (MTO) are held by their authors. Items appearing in MTO may be saved and stored in electronic or paper form, and may be shared among individuals for purposes of scholarly research or discussion, but may not be republished in any form, electronic or print, without prior, written permission from the author(s), and advance notification of the editors of MTO.

[2] Any redistributed form of items published in MTO must include the following information in a form appropriate to the medium in which the items are to appear:

This item appeared in Music Theory Online in [VOLUME #, ISSUE #] on [DAY/MONTH/YEAR]. It was authored by [FULL NAME, EMAIL ADDRESS], with whose written permission it is reprinted here.

[3] Libraries may archive issues of MTO in electronic or paper form for public access so long as each issue is stored in its entirety, and no access fee is charged. Exceptions to these requirements must be approved in writing by the editors of MTO, who will act in accordance with the decisions of the Society for Music Theory.

This document and all portions thereof are protected by U.S. and international copyright laws. Material contained herein may be copied and/or distributed for research purposes only.