II.   THE MELODIC FORM

(a) The types of melody.

W. FISCHER has made some important contributions to the study of form in the classics. In what follows I propose to assume and make use of the ideas he has advanced; so it may be as well to set them forth here briefly, in his own words. He distinguishes two types of melody - the song type (S-type) and the continuation type (C-type). Concerning the former, he says (p.25); "The antecedent of the period is a group of two phrases (α and β); in the subsequent these are repeated in their entirety, either unaltered or with modification of the second part (β). In every case α reappears note for note." (Ex.60)

And concerning the C-type (p.29): "After an antecedent with perfect or imperfect cadence, there follows a modulating 'continuation', where materials may or may not be related to what precedes it, and which consists of one or more consecutive sequences; frequently a third group acts as a 'conclusion' or 'epilogue' to the whole". According to Fischer, Ex.61 illustrates this type "in its highest perfection". A+b constitutes the antecedent, c is the continuation with sequences, d is the conclusion.

Fischer goes on to state (p.62): "In the neo-classical style, melodies, either of the S-type or closely allied to it as regards their form, are fitted into the frame provided by the baroque C-type." Or, somewhat differently expressed (p.52): "The C-type develops into the form of exposition of the sonata movement, with exclusion, of course, of the second section." Fischer (p.51) instances a theme by Pergolesi as an example of the transition, "where the first subject of a melody framed on the C-type is constructed on the S-type". Space will not admit of my dealing further with the matter here; I shall discuss the importance of these sections more fully in a later part of the book.

The definition of the two types of melody shows that repetitions and sequences fulfil an important role. And in Chapter I we have already dealt with these structures. So let us now consider the groups of alterations and those cases which have to do with the formation and arrangement of repetitions and sequences.

(b) Threefold repetition.

Gal (p.68) has shown how, in the episodic theme of the 'Coriolanus' overture, a four-bar group is formed by threefold repetition, and he has drawn attention to the logical way in which this is introduced. Later (pp. 111-12) he quotes similar instances of groups used in the working-out. As a matter of fact, it is possible to demonstrate this threefold repetition everywhere throughout the structure both of theme and of movement, and in their larger as well as in their more detailed aspects.

Before I proceed to this systematically, I should like to show, in one developmental series, how Beethoven arrived at the formation of themes of this kind. Nottebohm gives a series of sketches for the first theme of the quartet Op.18 No.1; let me reproduce some of these, together with the final reading.

62 b contains a threefold repetition, but it is very 'short-winded'.

62 c reverts to a sort of S-type with a fourfold repetition of the first motive.

After much experimenting, the final form 62 d shows the threefold repetition of the opening motive with the construction |: 2 :| + 2 + 2 (see abbreviations), in which, however, the third two-bar group is an intensified repetition of the first two.

It is not very easy to decide as to which type of melody this finally belongs. There can be no question of an eight-bar S-type, for the second two bars in no way correspond to the fourth two. The continuation (see Ex.109) suggests that we have here an eight-bar antecedent of a C-type.

There are many other instances of the kind. The scherzo of Sonata Op.14 No.2 shows three-fold repetition of the same short motive. In other respects this is a clear case of an S8-type (see abbreviations). It is obvious that a melodic structure of this kind must affect the whole movement; and, as a matter of fact, the threefold repetition reappears again and again.

A similar example (Ex.64) is the theme of the finale in Op.10 No.1; here also we have to do with an S8.

But this is not true of the second theme in the same movement (Ex.65); this consists of two motives in two-bar structure, each of which is repeated three times, though with changes of form.

Two-bar motives of this kind arranged in threefold sequence appear also in the theme of Sonata Op.2 No.2 (Ex.66); when it comes in for the third time; the motive is reduced to its elementary form and then carries on. This gives a C8 with the structure 2 x 2 (antecedent) + 4 x 1 (continuation), and having the motives of the sequences connected one with the other.

The second theme of the same movement (Ex.67) shows a threefold repetition of motives written in four and six bars. This gives a C-type with the structure 2 x 4(α) + 3(α) + 3 x 2 + 1.

Among many other examples I shall mention only the trio in Op.10 No.2 (Ex.68). The structure in this is clearly 3 x 4 + 4; accordingly, it can scarcely be regarded as an S, but rather as a special case of a C16. The same melodic structure that we find in the antecedents of the first and second themes also appears in the episodic sections and in the subsequents;

likewise, in the consequent in Sonata Op.7 (Ex.69a), we get the consequent of a C12:

and in that in the Sonata Op.106 (Ex.69b) the consequent of a C8: Gal pointed this out (Gal p.113) and he makes it very clear from the instances he gives. 'In his mature years Beethoven feels this threefold repetition to be redundant.' If this remark of Gal's is true of the threefold repetitions in the working-out, it also holds good of the thematic construction. In the compositions up to Op.30-31 such repetitions are very common; after that they occur only now and then, and in an especially happy form. We find them, for instance, in the theme of Op.106 (Ex.69b) and in the theme for variations in Op.109 (Ex.26); or again, in scherzi and similar movements, such as the presto in Op.130, where we have an S8 with the structure 3 x 1(α) + 1 + 3 x 1(α) + 1, or in the scherzando in Op.127, where the form is the same. In the presto in Op.74 we get a C8 running thus: 3 x 1(α) + 1 + 3 x 1(β) + 1 (Ex.36); the sketch makes the original intention still clearer.

This melodic structure has certain disadvantages; it produces clear-cut caesurae at every two bars, which become inconspicuous only when the tempo gets rapid. In his early work Beethoven was satisfied with this structure; later on he used it only where it could do no harm. In the threefold repetition in thematic structure we have one of the stylistic features characteristic of Beethoven as a young man.

We learn something more, however, from these considerations. The S-type is usually in eight bars, and, so to speak, square; this follows inevitably from its origin in folk-song and dance tunes (Fischer p.29). But in Beethoven the C-types of melody also have the same number of bars.

This would seem to justify Riemann's theory (Riemann.I p.198) that every melody is constructed in this form: the rule being 'that the continued contrasting of units of the same magnitude, the first answered by a second of the same order, builds the foundation of the structure of musical form'. In Exs. 62d, 68, and 69b there can be no question of this. The last part is set in contrast to the three first, which just suffice to balance it in spite of the greater number of bars they contain; conversely, the last part, in spite of its brevity, is able to intensify the effect through its very unlikeness, both rhythmical and melodic, to the other two. How the regular C-type is arrived at in the case of Beethoven's music will become still clearer as we proceed; it is already obvious that its structure is quite different from the S-type: 'the first subject is not answered by a second subject of the same order.' Sondheimer's 'proof that a square dance-rhythm is not the basis of modern instrumental music' seems to me to find support in the case of Beethoven through our discovery of this principle in his thematic construction.

Where else do we find melodic construction with threefold repetition of the motive? Mozart uses it very seldom, and then usually in episodic sections and in unimportant passages, such as occur, for instance, in the pianoforte sonata in C major (K.V. 309).

The theme of the C major quartet (K.V. 465) is one of the few melodies that show it.

In Haydn's music it is still rarer. Jalowetz, however, has pointed out the relation between the melody in Beethoven's early music and that of Philipp Emanuel Bach; and he has picked out a theme by that composer as counter-theme to the last section of Op.7 (pp.457-8). And this theme of Ph. E. Bach shows clearly what is far less obvious in Beethoven, i.e. the threefold repetition. As a matter of fact, this structure appears throughout Philipp Emanuel Bach's 'Sonaten für Kenner und Liebhaber'. It also occurs in the final allegro of sonata I in rondo II (Ex.78) in the second collection, in the first allegro and in the allegro di molto of sonata I, and in the allegro assai of sonata III - all of which are in the third collection; many other instances are to be found in the other collections. Here therefore we have a new feature common to both artists, over and above the one pointed out by Jalowetz and Oppel. For I consider that we are not dealing here with a conscious dependance of one artist on another; and an unconscious tranference of formal features of this sort is explicable only when there is common ground in the character of the music itself. I consider this condition very well exemplified in the case of these two artists.

As we have seen, threefold repetition served for construction of the theme; there are other cases where it is employed to build up a movement or a part of a movement. Gal goes so far as to interpret these threefold groups in the working-out as an attempt to 'expand the form' (p.113). In simple song-melodies we constantly find the threefold repetitions of a phrase, and especially of its first half; in the same way, it is very common to find a whole melody repeated three times in slow movements and in rondos. An instance of this occurs in the third movement of the serenade Op.8, where two melodies - adagio - allegro molto - adagio - allegro molto - adagio - appear alternatively; and the slow movement of the violin sonata Op.96 brings in the main theme three times. This can scarcely be regarded as a peculiarity of Beethoven's.

It is another matter altogether when we proceed to follow out the development of this idea in cases where the form has to be expanded in movements of minuet or scherzo character occurring in the works between Op.70 and Op.100 roughly. Here we have not so much a special feature of the scherzo, in Becking's sense, as a middle section, We find a transition to this in the scherzo of the pianoforte trio Op.97. The first part of the scherzo theme consists of 16 = 2 x 8 bars; then follows the second part; and to this is immediately appended the melody of the trio plus the first part. The repetition-signs show that all this is to be played twice; then follows the second part of the scherzo, and after that comes the coda, which gives us the following structure:
   Part I   |: Part II + Trio + Part I :|   Part II + Coda
   |---- Scherzo ----|
and this means that the scherzo melody occurs three times. This form occurs in full in the presto of Op.74, in the Presto of the seventh symphony, Op.92, in the trio portion of the allegro assai vivace ma serioso in Op.95 (though with fundamental changes), and then finally in the presto of Op.131. There have always been minuets and other dance forms with two alternating passages in which the main melody is repeated three times. But to expand this form with an episode is undoubtedly one of Beethoven's peculiarities as a composer; and even in his work it occurs only for a relatively short period. Later on, Schumann revived it in some degree, when, as in the second symphony, the pianoforte quintet and the third pianoforte trio, he expanded the several parts; but he made a departure by introducing two trios. This is an instance of the way in which a composer of the romantic school dealt with and modified the development introduced by Beethoven; in Beethoven's own work the form is expanded, but the thematic material remains unaltered; in Schumann the material is expanded, and it is the form that remains unaltered. Later on, as in the molto vivace of the ninth Symphony, we get, however, some expansion of the theme, in which fugato plays an important part.

In his earlier work Beethoven effected this expansion in other ways, as can be seen by comparing the octet for wind-instruments Op.103 with the string quintet Op.4.   Altmann has pointed out (compare also Orel) that the quintet is not a mere re-instrumentation of the octet, but has been rewritten and expanded from first to last. The final movement of 223 bars in the octet, for instance, is extended in the quintet to 418 bars. This movement can be divided up into some four sections. In the following table the comparison is brought out:

(c) Avoidance of sequences and uniform rhythms; characteristic relations between the melodic types, and their changes.

Writers have frequently laid stress on the importance of sequences in the formation of Beethoven's themes and elaborations generally. In addition to the passages already mentioned (in a certain measure the threefold repetition also belongs to this category). I shall refer to the treatment of individual themes in the third and fifth symphonies pointed out by Schmitz (pp.79, 84). I have already made it clear that this threefold repetition is a feature of Beethoven's early style.

There are actually instances where melodies, at first constructed on this principle, are altered so that the essential melody is retained but the repetition is avoided. The theme of the first march for four hands Op.45 No.1 shows in the final reading a substitution of a sequence for the third repetition in the sketch.

And I may also cite here as interesting the alteration in the theme for variations from Op.18. V (Ex.8) already referred to. The sketch shows threefold repetition, but at the same time a marked S-type: i.e. it shows liaisons of two kinds, as is shown by the following table: This gives the whole thing a rhythmical and melodic monotony which is avoided in the final theme, a pure S-type.

This association of repetition and sequences in the forming of a theme is a kind of logical construction from a motive provided by the imagination. Out of such a motive the construction promptly develops a theme, which has then to be modified in consequence of its too regular rhythm or of the predominance of sequences. I propose to deal more fully later on with the psycological aspect of this process: here I shall merely give a few more examples.

In Ex.6 we saw sketches for the final theme of Op.132 ; Ex.73 shows other stages in the development to the final form; and we see in these how the uniformity of the sketches is increasingly avoided. We shall return to the question of the theme later on.

As we may judge from the sketches, the beautiful middle section of the adagio of the ninth symphony also suffered from too great uniformity of rhythm, a fault that was soon corrected by alteration of the melody in the third bar, together with addition of an up-beat and syncopation.

At the beginning of the second part of the march Op.45 No.1 we see how a series of short sequences is avoided through rhythmical and melodic changes in the first segment as contrasted with those that follow.

Among other instances of the kind may be mentioned the theme of the last movement in Op.18 No.1 ; in the sketch this begins with a half-bar sequence which is avoided in the final form. In this example we see a number of things to which I have already drawn attention. In the sketch, the highest point of the first motive lies in a very unfavourable place; but in the final reading this is rectified. The apex of the theme comes early, so a repetition immediately follows; bar 7 contains a second apex, the way for which is carefully prepared. The continuation (we are dealing here with a C8) is formed by threefold repetition, it is possible that the fifth bar in the sketch suggested the alteration. We should also consider here the scherzo theme of the cello sonata Op.69 (Ex.35), where the threefold repetition is deliberately avoided.

Finally, let me mention certain cases which do not concern the theme but other 'complexes'. The end of the first part of the bagatelle Op.126 No.2 was sketched out by Beethoven as we see it in Ex.75a: and our admiration must be roused by the skill with which he avoids the sequences in the final reading Ex.75b.

In the sketch for the first bridge-passage from the first movement of the Eroica there is a triplication with a repetition of notes; then follows an ascent by means of a sequence in seconds, repeated three times. In the final form this is greatly abbreviated; it is written almost without sequence and in the manner of a lively up-beat. Finally, there is the instance, already cited in Ex.16, where the simple sequence is masked by the aid of lively up-beats formed in various ways.

In other groups showing alteration of the structure of the sequences we see the development of perfectly definite types of melody. The essential feature in these types lies in the relations between the several parts of the melody. The example on p.47 shows what these are. The way in which they are grouped is characteristic of each type. In the S-type they follow in successive groups, the second linked to the first (H. Riemann); isolated parts of the melody, such as bars 1-2 an 5-6, are alike, both as regards motive and rhythm (Ex.60). The C-type is different; the successive parts are connected by virtue of sequences and repetitions. Fischer has shown that 'where the antecedent is made up of a number of segments, the motives of the sequences in the continuation are always shorter; as soon as the continuation begins, we get a contraction of the rhythm'. In what follows I shall refer to this phenomenon as 'sequence-contraction'. I consider that what is essential in shaping the form is to order the relations of the final melody in such a way that they suffice to unify it and make it hold together; but they must not be so numerous that the independence of the several parts is seriously imperilled: for this leads to the form's becoming unrecognisable. Ex.8 illustrates this. It is possible to find in the sketches a number of groups in which we can see a unified type of melody develop out of a condition in which there was a superfluity of liaisons consisting of sequences and regular rhythms. The theme from Op.132 (Exs. 6 and 73), to which I have aready referred so frequently, belongs here. From the first drafts it is not possible to ascribe it to any particularity of melody, either on account of, or in spite of, the rhythmic liaisons between bar and bar (Ex.6); it has a certain unity, but is not well put together. In its second stage (which, however, may possibly not belong to this developmental series) it is a C-type; there is a four-bar antecedent with two-bar liaisons, and at the end we end a Sequence-contraction. The next sketch (Ex.73b) has none of these organising liaisons; it is to be regarded merely as an eight-bar antecedent of a C-type which also has a course of eight bars. In the final form (Ex.73c), which is distinguished from the others by something additional ( See Chapter II g), liaisons are again established between bars 1, 3 and 6 by means of similarity in the rhythmical treatment.

The theme of the second movement in Op.14 No.1 is badly held together in its sketched form (Ex.77a). It begins with triplication of one bar; then follows a two-bar repetition. Obviously this is very faulty architecture. And in the final form it is changed. Here the repetition of the first bar is cut down and that of the two-bar part is dropped out altogether; the theme seems to be written in half-bars, and the end-result is an indubitable S-type of 16 = 8 bars, with liasons characteristic of that type. Just as in the previous example, this S is a sort of antecedent in a larger C-type.

The adagio theme of Op.59 No.1 had originally only six bars, the fifth with short sequences; the order of the liaisons was not clear. Out of this developed a C8 with a four-bar antecedent and four-bar continuation; the rhythmical connection between bars 1-2 and 5-6 is too slight for an S. And here we see something that we shall frequently meet with again, i.e. the birth of a regular eight-bar theme out of one of about six bars ('irregular' in Riemann's sense). But, as was pointed out above, its genesis shows clearly that the structural form here is that of an C- and not of a S-type. This form and the same organization also appear in the episodic passages.

Ex.79a gives the episodic bars in the first theme of the concerto Op.15 as they appear in the sketch: there is an obvious triplication, and bar 7 shows reference to bars 1, 3 and 5; but we find no development of any kind. In the final form this superfluity of liaisons is avoided, and instead we get a regular S-type, in which the essential theme is preserved.

The allegro theme of the Sonata Op.101 also furnishes a beautiful instance of the way in which liaisons of this kind are changed into a C8. The original draft is short-winded and without definite form; from it is evolved a fine example of a C8. The antecedent has four bars with the structure 2x2; the continuation also has four bars showing sequence-contraction, 2 x 1 + 2. Here again we see what importance Beethoven attached to getting the form clear; for in the final solutions thematic unity is always associated with an arrangement of the liaisons that leads to greater intelligibility.

Op.59 No.2 and the first movement of the eighth symphony contain two very instructive examples. In Ex.23 I gave three sketches for the presto of this work. Ex.23a is an S-type that shows a very special peculiarity. Elsewhere (Zeitschrift für Musikwissenschaft) I have shown that it is not justifiable to apply the concept of 'symmetry' to the normal S-type; but here symmetry is actually produced, and might be reproduced diagrammatically somewhat as follows; The second sketch (Ex.23b) is even more unusual; in this it is impossible to recognise either the S- or the C-type.

And we find that Beethoven also rejected this arrangement of the liaisons. What did he substitute for it? An eight-bar C-type with a four-bar antecedent (||: 2 :||) followed by a continuation of four bars showing sequence-contraction (Ex.81).

It would appear that the symmetrical liaisons did not satisfy Beethoven's architectural sense; the episodic theme of the first movement in the eighth symphony seems to corroborate this impression. In the sketch for this the line to the apex is quite symmetrical. Both the antecedent and subsequent are formed by amplification; and, since the motives correspond, this makes the whole look as though it belonged to the S-type. In the final form the same construction is kept, but the two motives are differently fashioned, so that the curve is different on its two sides. We get more the impression of a C8-type: which shows what a narrow line separates the two types of melody when it comes to the extreme cases.

In the article referred to, I showed that it is possible to make music strictly symmetrical by technical means without our feeling corresponding to this in any way. The way in which these two almost symmetrical themes were altered, and the fact that really symmetrical melodies seldom, if ever occur, seen to justify my opinion and to give it the support of a very important witness, i.e. Beethoven himself.

Lorenz has shown that in the arrangement of Wagner's melodies there is a plan - a b c b a - and this he calls the 'perfect curve' (a, b, and c stand for the several melodies). But the fact that a structure so simple as this has escaped detection for so long is, to my way of thinking, just so much the more proof that no acoustic or aesthetic equivalent for it exists.

In this connection the change made in the scherzo theme of Op.106 is instructive. It has often been observed that the melody consists of seven bars. The sketch (Ex.24a) is a C8 with a beautiful sequence-contraction and the construction 2 x 2 (anteced.) + 3 x 1 + 1 (contin.) . But the same movement is also held together by liaisons which occur either between bars, by means of their common rhythmical structure , or by the ascending and descending seconds in bars 2, 4 | 5, 7. In the final form, it is true, these liaisons are cut out, the repeated motive is shortened by one bar (probably on account of the early apex) and the triplication is added. The concluding bar should really come then, but this does not appear in the original form. lt was omitted, and in its place there was added a further triplication with an appropriate final bar. Becking (p.141) has already pointed out the importance of this irregularity with regard to the scherzo character. This seven-bar melody is undoubtedly a peculiarity, and it can never be understood as an S-type, but as a C-type in conjunction with threefold repetition. An example like this shows that Beethoven was aware of the superabundance of liaisons and sequences; the quick tempo (double bars, according to Becking) makes the rhythmical structure from bar to bar less conspicuous.

Similar experiments - though less successful in the final form, it is true - prepared the way for the theme for variations in the quartet Op.74. A first sketch (Ex.83a) shows clearly the structure of the threefold repetition 3 x 2+2, or, where in the third the change appears too extensive, the eight-bar C-type 2 x 2 + 2 x 1 + 2; here the same motive appears six times.

In Ex.83b an attempt is made to improve on this by a change in bar 3, without further justification.

In the final reading the two-bar sequences are cut out. Does this not involve a loss of structural stability? It is clear that Beethoven recognised that the sequences of the original form were to some degree logical, but the construction finally adopted is not quite sure, nor quite intelligible.

That the formation of such sequences may be essential to the musical expression is shown by the developmental stages of the melody 'In des Lebens Frühlingstagen' from 'Leonore'. The sketches in Ex.84 make it clear that Beethoven was trying to use as much as possible descending seconds, intervals that are well known to produce a soft and sighing effect (P. Mies, 'Über die Tonmalerei'), further aided here by the suspension. (See p.36) But to get shape into this involved difficulties. The first sketch (a) has still got two ascending seconds, one third, and one ground-note; sketch (b) has a ragged beginning and is without liaisons; sketch (C) has been smoothed down but is otherwise much the same as (b). Sketch (d) is the first to contain definite Sequences, but it also has an up-beat third and fourth. It is only in the following draft that Beethoven attains his object; except for one fourth, there are only descending seconds and ground-notes, and, as a consequence of the early apex, there are very beautiful sequences and repetitions. In the final form (f) the second apex is further emphasised by a grace-note, and a descending second is introduced; by these means the intended expression is achieved. I consider that in the works of the great masters we do not find excessive divorce between form and content: nothing short of the right form will release the desired content; the expression sought will reveal itself only when the form adequate to it has been found. The two are indissolubly connected. (E. Utitz comes to similar conclusions in 'Zum Schaffen des Künsters')

In the foregoing paragraphs I have dealt mainly with the excision of superfluous liaisons conditioned by the rhythmic structure or by sequences. I shall now refer to some cases in which the liaisons depend on similarities in the melodic structure. It is impossible, of course, to draw a hard and fast line between these and the preceding. All I am concerned to show is that Beethoven finally selected, from out the superabundant liaisons, just those which were characteristic of the particular type of melody adopted in the final reading.

And in this connection there is an interesting case (Ex.85) quoted by Gal (p.94) from the trio in the second symphony, where the melodic liaison in bars 5 and 7 is weakened, so that there remain only those belonging to the S-type.

The theme of the trio in the eighth symphony was originally written with a threefold repetition on the plan 3 x 2 x 2 (Ex.4). The next sketch shows its transformation into an S-type, but still with the first and third bars alike: which is an unnecessary relationship. In the final reading this also is eliminated, and the consequent is given more variety.

We find this latter improvement elsewhere - for instance, in the trio of the Eroica, which at first was sketched out as the simplest sort of S-type (Ex.86).

In the theme of the last movement of the quartet Op.18 VI the form given to the liaisons is altered very extensively. The first experiment (Ex.87a) makes use of threefold repetition; but the result is unsatisfactory and lacks unity.
The second sketch (Ex.87b) also shows the triplication, but the repetitions are linked together into an S-type; thematically, they are derived from the final motive of the first.
In the next attempt (Ex.87c) a true S-type is attained, but there are even more liaisons in it.
The theme when finished (Ex.87d) is very important as regards its form. It is an S 16: the antecedent and consequent have different cadences, but each is a C8. The result is a mixed type, such as we shall frequently encounter again. I propose to describe them as S (C) (see abbreviations). The structure of the whole theme then is as follows:

   2 x 2  +  2 x 1 + 2  +  2 x 2   +  2 x 1 + 2
antecedent continuation, antecedent continuation   C-types
     Antecedent CF7         Consequent CF7B6  of S-type cadences

The above examples should suffice. We see in them clearly that Beethoven was striving after an intelligible and characteristic thematic form; but at the same time it is apparent that the completed form was not a flash of inspiration that came at the first moment, but the result of retaining a certain rhythmical and thematic structure throughout the many changes to which the work was subjected.

In settling what are the liaisons characteristic of a melodic type, we may find that the reverse process occurs: it may be found necessary to introduce liaisons that were lacking in the sketch. As a matter of fact, this happens more seldom. In the sketch (Ex.88a) the theme of the minuet in Op.59 No.3 is not of any definite type, but in the final reading (Ex.88c) it becomes a typical S8.

The theme of the second movement of the ninth symphony was made up of two fugue themes dating from the years 1815 and 1817 (Ex.89a, 89b).

The beginning of a longer sketch (Ex.89c) shows a similar mixture.

Little by little, the raggedness of the beginning is smoothed away, the quavers disappear, and with their dissapperarance the melodic line hecomes better unified (Ex.89d).

From the sketch (Ex.89e) we may perhaps conclude that the idea of 'Ritmo di tre battute' was the germ from which this unification arose.

The theme of the second movement of the fifth symphony had to pass through the same process of transformation. Ex.90a gives a sketch with the superscription 'Andante quasi menuetto'. The sequence is obvious; the dotted rhythm in the writing in quavers fits the minuet character.
Another sketch (Ex.90b) shows these dotted rhythms only in the last bars, i.e. in the continuation; and, apart from the altered up-beat, the sequence seems to me less clear; the descending third, D flat - B flat, in bar 3 does not correspond so closely to the C - A flat in bar 1.
A further sketch (Ex.90c) shows that the dots were not merely forgotten.
In the finished work (Ex.90d) the dotted rhythms appear again, and in this way the descending third in bar 3 is once more brought into prominence. The result is a definite C8 with the structure
    2x2 (antecedent) + 2x1 + 2 (continuation)

The sketch for the minuet melody in the trio Op.1 No.3 (Ex.91a) is also without characteristic liaisons.

In the final form (Ex.91b) the first and second bars of the sketch are cut out, and as third bar a motive is introduced that corresponds with bars 7-8 in the sketch.

The outcome of these changes is a regular S-type with a rest in place of the eighth bar. This S-type is the antecedent of a C-type, and the whole theme has accordingly the structure of a C(S)-type (see abbreviations), with the structure
    2+2+2+1+fermata (antecedent) + 3x1 + 3x⅓ + 1 (continuation)

The adagio theme of Sonata Op.96 is also without unity of form in the sketch (Ex.92a). The relation between bars 5-6 and 1-2 is too slight for an S-type; conversely, the similarity between bars 5-6 and 7-8 is out of keeping, and the development necessary for a C-type is lacking.

The finished work, on the other hand (Ex.92b), is definitely of the S-type; bars 1-2 and 5-6 are rhythmically alike, and the repetitions have dropped out.

(d) Sequence-contractions.

I have already pointed out (p.61) that I agree with Fischer on the importance of sequence-contractions as a term of melodic liaison. The relation of the length of sequence in the antecedent to that in the consequent, and the working-out of the contraction, are matters of vital importance to the elasticity of the theme in the C-type. In one group of alterations, where the theme may originally be of the S-type, the characteristic feature is the development to a C8, C16, etc.

The last movement of the trio Op.97 had in the sketch (Ex.93a) a C8 as its first part, with the structure 2x2 + 4; this was followed by an S8 as the second part. In a later sketch this second part has already undergone contraction.

From this, the final reading (Ex.93b) proceeds logically on the plan 3x1 + 3x1 + 2. The whole thing, accordingly, is a C, in which both the antecedent and the continuation are also built on that type.

The small subsidiary half-bar sequences in the continuation are important to the development, which is completely lacking in the sketch, where the two types are merely set side by side. The special character of the form is in no way affected by the fact that the repetitions are varied in the two parts although the harmonization is the same.

The middle portion of the funeral march in Op.26 displays well the intensity and force which, in certain circumstances, the contracted C-type may assume. The sketch has an antecedent (Ex.94a) of four bars, which may be regarded as a sort of S-type.

In the finished work (Ex.94b) this somewhat trivial regularity is replaced by a C4 with wonderfully effective contractions. It is built as follows:
    2x1 (anteced.) + 2x½ + 3x¼ (contin.) + ¼ (conseq.)
This effect is much employed in episodic sections.

In the minuet theme of Op.59 No.3, the first part of which appeared in Ex.88, the interlude originally began with a phrase twice repeated (Ex.95a).

In a later sketch (Ex.95b) it began with a very short contraction. The theme still lacked development to the repetition.

This appeared only in the final form (Ex.95c), where a series of contractions lead up to the repetition with admirable inevitability.

As regards the opening theme of the Appassionata (Op.57), with sixteen bars built on the C-type, I have only to bring together sketch (Ex.96a) and the final reading, (Ex.96b) and at once we must be struck by the steadiness of the final form, in contrast to the abruptness of the contraction in the sketch. It should be observed also that the C-type finally evolved is again in 16 bars.

The development in the presto of the seventh symphony is equally clear. In its first form (Ex.97a) this is a rather ineffective S-type with a contractive curtain:

later on (Ex.97b) it consists entirely of two-bar segments without connection or development;

finally (Ex.97c), it becomes a boldly conceived type with the following structure:
    2 (curtain) + 4x2 (anteced.) + 2x2 + 2x1 (contin.) + 3x2 + 2 (conclus.)
which is 24 bars. It is clear that Beethoven experienced some difficulty in developing the form; but create it he did by sheer hard work, guided, whether consciously or not, by certain definite principles.

In these last examples we find the development moving towards more wide-embracing thematic complexes. Since Beethoven now preferred to begin themes with sequences, and since such structures tend to involve contraction, a type was produced that begins with a C8 and then repeats this with a change in the second part. The whole theme, therefore, is of the S-type, in which either the antecedent or the consequent, or both, are of the C-type. I have already referred to an S (C) of this kind from Op.18. No.6 (Ex.87). In Ex.91 from Op.1 No.3 we get the converse type C (S). Both were very common in Beethoven's works, and both tend to increase the complexity of the theme (the thematic complex).

I shall mention a few more instances from the S (C) group. As it appears in the sketch (Ex.98a), the rondo theme of Sonata Op.24 has a continuation that is only very slightly developed (ed: key signature ? 6# ?),

whereas in the final form (Ex.98b) it progresses from half-bar to quarter-bar sequences. The antecedent is a C8; by means of repetition, with altered and intensified cadence, an S (C) type of 18 bars is produced.

There is a similar development in the theme of the adagio in Sonata Op.10 No.1. The sketch shows an eight-bar melody, which, to judge from the similarity in the rhythmical structure, was conceived on the basis of threefold repetition; but it was not to be repeated. The variant introduced a sequence-contraction; but the apex, A flat, was left in an unfavourable position.

The final form steers clear of this defect and begins the continuation at once with the contraction and repeats the antecedent with a different cadence, so that the whole become an S (C) type of 16 bars. The fact that most of the S (C) types have an even number of bars does not follow as a matter of course; it is merely a consequence of what I have often mentioned already, i.e. the regular number of bars in the types of which it is composed.

Müller-Reuter states (with Czerny as his authority) that in the conclusion to the first part of the Eroica (ed: end of the exposition, bars 132-155) there were originally two more bars "which complete the rhythmical framework of four bars". It is quite possible that the answer to this problem is bound up with Beethoven's attempt to shorten sequences and so to abbreviate the rhythmical structure generally. The two bars referred to are marked by a X in Ex.100a.

Ex.100b gives the whole passage. It is built up as follows:
  4+4+4+4+2+2+2+2
Interpolation of the two bars would make it:
  4+4+4+4+2+4+2+2
Accordingly, the present solution seems to be the better and more justifiable.

This suggests an interesting inquiry into the rhythmical construction of Beethoven's compositions in general, i.e. the way in which the larger parts are put together. But obviously such an inquiry would be beyond the scope of the present work.

All the foregoing examples show how the production of form may depend on the cutting out of liaisons, rhythms, sequences, along with the introduction of thematic connexions and sequence-contractions. They make us see the advantage of considering the question from at new angle, and, in place of studying the means by which the remodelling is effected, of demonstrating rather how the first ideas as they appear in the sketches are related to the final form they assume in the finished works, and to what types these belong.

(e) Transformation of the types.

I propose to deal first with those themes which develop into an S-type. In Ex.92 I have already cited the adagio theme of Sonata Op.96, in which an S is derived from a melody with disconnected liaisons. In the theme of the theme of the Eroica (Ex.86) a varied S-type is developed from a simple one.

If we compare the sketch (Ex.101a) with the final form of the adagio theme in Op.30 No.2, we see a marvellous unfolding from out a stereotyped form.

In the final reading (Ex.101b) the S-type liaisons almost disappear, or are, at any rate, profoundly modified - at this stage they cannot be completely destroyed. This theme shows Beethoven at his most typical; in a later section I shall discuss more fully the method by which the transformation is effected. The characteristic feature in this kind of melodic construction is the remarkable emotional breadth.

It is not common to find hesitancy such as we get in the scherzo theme of Op.97. Ex.102a gives the first attempt written as the simplest sort of S-type, with a very monotonous rhythmical structure. But the composer has found the right rhythm; and the way in which he works up this theme to its prefect form seems, at this stage of the development, to be the result of logical seriation rather than of imagination.

A second sketch (Ex.102b) written in the C-type has sequences in the consequent, but none in the antecedent.

A third (Ex.102c) shows the C-type with sequence-contraction, and in this draft the final motive has been found, in addition to the rhythm.

The final form (Ex.102d) is a varied S 16.

We had another instance of this transformation of the C- to the S-type in Ex.4 from the trio in the eighth symphony (see p.7). This practically exhausts the list of cases forming the group of changes in form that lead to an S-type. Moreover, it is obvious that the theme with the liaisons characteristic of the C-type cannot straightway be turned into the S-type. For the liaisons in the latter are more fixed and wider in their embrace than in the former; and in the S-type the antecedent and consequent are more closely connected in their first segments.

The reverse change-of the S-type into the C-type is therefore more common, especially as the liaisons in the consequent are easily modified by means of sequences.

The episodic section in the Sonata Op.30 No.2 (bars 28-36) is a C8 with sequences in the third, fifth, and seventh bars; the original form was an S8.

The theme of the minuet in Op.18. V (Ex.37) was of a varied S-type in the sketch; but in the final form it is a C 12 with the structure:
  4 (anteced.) + 2 x 2 (contin.) + 4 (conseq.)

The theme of the presto in the seventh Symphony already referred to (Ex.97) also shows this development.

I shall give as my last example the first theme from the sonatina Op.79. The first sketch (Ex.104a) is of the S-type;

in a later sketch (Ex.104b) the theme is changed, without characteristic liaisons being introduced.

Finally (Ex.104c) an S8 appears, in which the continuation is added to the antecedent in the form of a sequence.

I shall now refer to some instructive instances in which the sketch belongs to the C-type. When we were considering the up-beat, I mentioned the eight-bar theme of Op.132, together with the sketches leading up to it (Ex.19).

The final shaping of this theme requires as little further explanation as does the growth of the first theme in the clarinet trio Op.11 (Ex.105).

From among the many important instances let me give a sketch for the fifth symphony along with its final form. The sketch consists entirely of sequence-like two-bar segments, of which only the last show contraction. The finished form has the structure:
  5 (curtain) + 2x4 (anteced.) + 3x2 (contin.) + 2 (conclus.) = 5 + C16

Schmitz (p.79) says of this passage, 'We might also imagine it scored as follows: (Ex.107) It is easy to imagine the guiding melodic line as played by the first violins alone, the sustained notes being held by the second violins, violas, and cellos. The imitations . . . as they appear in the actual score, do nothing to alter the rhythmic symmetry of the structure,'

But I consider that the regular one-bar rhythm would become too insistent, and would drown the effect of the four-bar sequences. The way in which the form is developed shows that the instrumentation is essential, and determines the structure here; it is not a secondary consideration as Schmitz suggests.

The opening theme of the Overture Op.138 (i.e. Leonore Overture No.1) was at first written in seven bars (Ex.108a). Nothing in the consequent corresponds to the sequence in the antecedent.

The final form retains the sequence and consists of eight bars. The retention of the first dotted minim throughout all the sketches shows that it was an essential part of the melody; the last note acts both as end and as beginning. Here again, we see the C-type expanded to eight bars.

I have repeatedly pointed out that in Beethoven's works the C-type is either 8 or 16 bars in length: and so, as regards the number of bars, it is broadly comparable with the S-type. The latter must of necessity have this number of bars, but the former is not limited in that way. The liaisons characteristic of it are quite different, and Fischer has thoroughly demonstrated the basis on which they rest. Riemann is certainly wrong in thinking that all themes have the S-type as their foundation. This explains, however, why his method of analysis so often does violence to the themes he is considering.

This connection that exists in Beethoven's music between the two essentially different types explains the important part played in his compositions by the mixed forms, S(C) and C(S), to which I have already referred. Both forms fall in readily with the four-bar structure. They are far wider in their embrace than are the simple S- and C-types. And it is one of the peculiar features of Beethoven's style that each part of a phrase shall encompass more than do those that precede it.

Let us first consider simple cases where, in an S-type, the antecedent or the consequent is, or both of them are, of the C-type; the result is a mixed type, S(C). We saw an instance of this in the finale theme of Op.18. VI (Ex.87) where both antecedent and consequent are C8.
In Ex.99 (Ex.99) likewise we get the development of the adagio theme in Op.101 to a 16-bar S(C) of the form 8+8.
In the theme of the rondo of Sonata Op.24 (Ex.98) the subsequent is lengthened so that the S(C) type has 8+10 = 18 bars.

The first theme in Op.18 No.1 is more complicated. The eight-bar antecedent is built up from a triplication (Ex.62). In the consequent the first four bars are retained and a sequence-contraction of eight bars is added, making it a C12. It is doubtful whether Schmitz (p.72) is right in saying that the following nine bars (bars 21-30) belong to the theme itself; they also contain sequence-contractions. From the repeat we can see that the first eight bars alone constitute a type of melody: it retains only bars 1-8 and then modulates for eleven bars after the manner of bars 21-29. I conclude from this that the theme ends with bar 20. The first draft (see Wedig) has the same thematic structure, except that there are nine modulating bars in the repeat; the sketches (Ex.62) show what difficulty Beethoven experienced in shaping this theme.

The bagatelle Op.126 No.3 furnishes a very beautiful development along these lines. In the sketch (Ex.110a) there are no characteristic liaisons whatsoever.

From this, by retention of the first four bars and elaboration of the bass motive in bars 10-11, the final form (Ex.110b) developed a C8.   With a four-bar antecedent and continuation; repetition with modified cadence made this into a sixteen-bar S (C). The postlude is not part of the theme: for the repetition,while varying the theme, fundamentally alters the consequent.

An instance of the converse case (where the whole theme is of the C-type, but its several parts - usually the antecedent - are of the S-type) is shown in the adagio theme of Op.18 No.2. In one sketch, it is true, we see much of what characterizes the final form; but the close juxtaposition of 2 + 3 bars and the very short non-thematic sequence-contractions in bar 5 prevent any feeling of the form coming through. The final draft, with its elaborate structure, is very different.
    1+2+1+2 (anteced.=S6) + ||:2:|| + ||:1:|| (contin.) + 2 (conclus.) = C14
Here, since the S-type appears cut down to six bars, the C-type has 14 bars instead of 16.

Even if we choose to reckon as part of the theme the twelve bars that follow, up to the intervening allegro, this would only mean that the conclusion had also become a C.

A theme in Op.131 shows great complexity of structure, although in the sketch (ex.112a) it is merely a simple C8.

In this case also the first four bars are retained and expanded to eight by repetition (ex.112b) . These eight bars are repeated with so much alteration in the instrumentation and the accompaniment, and with such different cadences, that we may speak here of an S16 as antecedent; then follows the 8-bar continuation with contraction. The sketch shows conclusively that these bars belong to the theme.

It may be appropriate here to trace the evolution of the first theme in the adagio of the ninth symphony. A preliminary sketch is of the C-type (Ex.113a), with an eight-bar antecedent whose articulations are but little marked.

The next sketch (Ex.113b) substitutes for this a definite S8 followed by a continuation.

These liaisons are still more schematically shown in a later experiment (Ex.113c), where the complete echoes are introduced only after the main sections.

In the final form (Ex.113d) the echoes are shortened but occur more frequently; they are no longer mere interpolations, but aid in the production of form. They constitute a kind of sequence, and so help out to the liaisons: for it is not easy to decide whether the final theme belongs to the S- or to the C-type. As a matter of fact, the two types are completely interwoven in the sketches. Here we are dealing with a theme that is characteristic of Beethoven's mature style; in a later section (Chap. II(g)) I shall attempt to explain the peculiarities that such themes illustrate.

As an example of an unusual development, I may mention the allegro theme of the overture Zur Namensfeier Op.115. In Ex.114a it is composed of seven bars - if an unwritten final note, E flat, be included.

It still keeps this number of bars in Ex.114b, but now contains sequences, the elements of the C-type.

The object of the two following experiments (Ex.114c and d) was to arrive at a different number of bars by means of sequences.

Ex.114e gives the result; the theme now runs to nine bars. What is unusual, however, is that an S 16 follows this C9, and is actually in melodic contrast to it. In Beethoven's work this occurs seldom (Gal p.68) and it is quite contradictory of Schmitz's principle of 'contrasting derivation'.

Finally (Ex.114f), the double bars are combined as simple 6/8 time. The irregularity is retained; the antecedent in the C-type has an apparent length of 5 bars, but as a matter of fact the number of bars is even, for the first is an up-beat or curtain; the high note always lies in the second part of the bar, as is shown by sf frequently being written above it in the overture; 114e is very different in this respect. Half-bars would therefore give the correct number of bars and the stress in the antecedent; in the consequent the reverse occurs.

All this may explain why this overture is not accounted one of Beethoven's best works. The process of getting it into shape cost him a great deal of labour, and even in the end he did not achieve a perfect result. The substitution of the S-type for the continuation is not really suitable.

Ex.91 from Op.1 No.3 gave a case in which the antecedent in a C(S) is an S, and this method of building up the theme gives essentially better results.

What we learn from these and similar examples be summarised as follows:

1. In addition to the simple S- and C-types, mixed types play an important part in Beethoven's work.
2. The S-type is the only one that is necesssarily limited to a length of 8, 16 etc. bars; but the C-type also is usually of this length.
3. This regular number of bars must therefore also appear in the mixed types.
4. Preference for the mixed types to enrich and expand the theme.
5. The work of remodelling themes proceeds mainly from the basis given by the motives determined from the outset.
6. Accordingly, establishment of the form and of the liaisons that control it must come at a later stage of the work, and only very rarely do these flash on the composer at the first moment in the way the thematic material does.
7. In addition to complete remodelling, the points already mentioned (such as up-beats (p.13 et seq.) and, in certain circumstances, the position of the melodic apex (p.27 et seq.) may be used in the production of form.

(f) Themes with an irregular number of bars.

Before going on to investigate a final characteristic displayed in Beethoven's themes, let me say a word concerning certain themes with an irregular number of bars. It would appear that in many cases special features conceal or counter-balance these irregularities; such examples accordingly serve to strengthen statement 2 above.

We can also see the derivation from the sketches. In Ex.111 we had a theme from Op.18 No.2, of the C(S) type and of 14 bars. The antecedent of the song has six bars; the sketch shows the development from four regular ones; the irregularity is scarcely perceptible. The minuet from the trio Op 1 No.3 (Ex.91) has as antecedent an S-type of seven bars; it is completed by the addition of a rest to the last bar and of five bars to the continuation that follows. The effect of the whole is to make the irregularity of the several parts less obvious. The original nal sketch had eight bars, and the same is true of the Scherzo theme in Op.106; whereas in the final form the number of bars is seven. I have already explained why the one bar was taken out (Ex.24 and p.67). The first theme of the quartet Op.18 No.5 (Ex.54) is also only apparently eleven bars long; the sketch had the structure 4 (curtain) +2+2+2+2 (S8). I have explained that Beethoven replaced the note-repetition by a swinging movement, and this involved shortening the curtain by a bar; the actual theme is an S8. The adagio of the ninth symphony (Ex.113) is full of repetitions, which involve a new irregularity in the number of bars, an irregularity that does not appear in many of the sketches. I have already drawn attention to the important part these echoes play in form-production; but they have another purpose, and the consideration of this leads me on to yet another feature in Beethoven's style.

(g) Melodic breadth and elimination of caesurae.

Wetzel in an account of Kurth's work 'Romantic harmony' says: 'Bach and Beethoven took great pains to achieve an increasingly perfect rhythmical organisation of the succession of sounds in their works. Kurth fails to recognize its development to subtlety from a very obvious foundation. He fails to realise . . . that Bach and Beethoven in their mature years composed melodies without definite limit, which overflowed the strophic limits, and that they were not in the least ashamed of these.' Throughout this chapter we have striven to follow 'these subtle developments to a perfect rhythmical structure' in the case of Beethoven, and, in so doing, to reveal the course he pursued and the laws that were established; for the liaisons characteristic of the several types of melody are nothing more or less than a rhythmical organization of this kind. Comparison of the sketches and the finished works would also have brought us to the conclusion expressed in Wetzel's second statement.

I have already analysed the remodelling undergone by the theme of the allegro appassionato in Op.132 (Exs. 6 and 73). The liaisons characteristic of the melodic types form sections separated by caesurae. Normally, the eight-bar S-type has one main caesura between the fourth and fifth bars; and between the second and third or sixth and seventh there are subsidiary caesurae. In the C-type all are distributed in accordance with its structure; sequences naturally tend to involve caesurae of this kind. Now let us consider the effect of such a theme in its final form. The literature on Beethoven is unanimous in its opinion. Helm (p.290) calls it 'one of the most impressive that he ever discovered'; Thayer (V p.269) writes: 'The melody of the movement is certainly one of the most beautiful that Beethoven ever wrote'.

Nothing of all this is apparent in the original sketch (Ex.6a); one bar follows the other with a regular rhythm. The sketch Ex.73b, on the other hand, has too few liaisons for the production of perfect form. The final reading (Ex.73c) has liaisons between bars 1 and 3, but it also carries over from Ex.73b bars 4-7, and, in so doing, acquires something essential; except for the caesura between bars 2 and 3, the melody flows on across all the caesurae. And the device that Beethoven employs here is characteristic; repeated notes, which in general he avoids (Chapter I(c)), are introduced in association with unusual rhythms (bars 4-5). In spite of its brevity, the melody seems to flow on without interruption.

The same thing may be seen in other examples. The bagatelle Op.126 No.1 was at first a varied type with marked caesurae in bars 4, 5, and 6.

Again in the final form (Ex.115b) we find these caesurae bridged by repeated notes and syncopated slurs. At the same time, the subsequent is so markedly altered that it loses its liaisons with the antecedent, and a C8 is produced.

I have already drawn attention to the filling-in of the rest in bar 2 of the slow movement in Sonata Op.90 (Ex.1 and p.5). Now we see that the reason for this was to get rid of the caesura in that bar. Concerning this particular instance Nottebohm remarks (N. II, p.366) that by the alteration of one note 'the melody is given a significance that it originally lacked'. The change is effected, however, not by this one note alone, but by the repeated note which bridges the caesura.

The theme for variations in Sonata Op.109 (Ex.26) illustrates the same thing. The sketch still shows clearly the main caesura between bars 4 and 5; in the final form this is bridged by a repeated note and an appoggiatura chord. The way in which this chord is written is not mere chance: it imposes a dynamic connection with the preceding part of the bar, and - let me once more emphasize this point - it puts the melodic apex in an effective position.

Even the delightfully flowing melody of the song 'Mit einem gemaltem Band' did not always move so smoothly. In the sketch (Ex.116a) it is clearly of the C-type, with the structure 2+2 (anteced) + 2x1 + 2x½ + 1 (contin) = C8. The caesurae in bars 2, 4, 5, and 6 are very conspicuous.

In the final form (Ex.116b), this is changed. Here the liaisons are weakened, the caesura in bar 4 is got rid of by repetition of a note over the same accompaniment, and that in bar 2 by removal of the rest.

The theme of the adagio in Sonata Op.101, which I have already mentioned (Ex.99), shows, even in the sketch, an attempt to eliminate caesurae in bars 2 and 4 by means of chromatic progressions that intensify the melody; in bar 6, on the contrary, the caesurae are obvious enough. In the final form these are bridged by sequence-contraction and the unexpected pause, followed by syncopation, on the premature melodic apex. We have here an early example of Beethoven's attempt in this direction, so it is not surprising to find it in a slow movement; for just these movements have always been felt to be especially characteristic of his style.

In the fourth bar of the adagio in Op.59 No.1 the harmonized suspensions (Exs. 52 and 78), by means of the tension they produce, bridge to some extent the marked caesurae in the sketch.

Here the bridging is made in a forward direction; but in the theme for variations in the Violin Sonata Op.30 No.1 the same result is achieved in a backward direction by means of an anticipated entry, as can be seen at a glance when one compares the sketch (Ex.117a) and the finished work (Ex.117b).

The sketch of the theme of the finale in Op.127 breaks off weakly: it contains only five bars, and is a late example of a theme produced by means of three-fold repetition. From the fourth bar in the sketch we get bars 4-8 of the final form, and an effect of even, uninterrupted flow. In contrast with the short-winded segments of the beginning, the breadth of the melody is twice as effective; from bar 3 onwards there is no trace left of any caesura.

The foregoing examples and the explanation that I have given of them justify the following conclusions:

1. Melodic breadth is a feature of Beethoven's style as shown in his later works, but it is also to be found in the early slow movements.
2. This breadth is etfected, in the main, by elimination of the caesurae characteristic of the particular type of melody.
3. For these he frequently substituted repetition of notes, syncopated liaisons and chromatic melodic progressions.
4. Melodic breadth may call for a certain reduction in the liaisons characteristic of the type, without their being rejected altogether.
5. Beethoven's tendency to introduce lively up-beats (already referred to in Chap. I (a)) may have originated in his attempt to give breadth to melody (Ex.1).

Having reached these conclusions through comparative study, we shall not find it difficult to substantiate them by reference to the finished works. A few examples of this will suffice. In the adagio of Sonata Op.2 No.1 (Ex.119) we have an S-type without the caesura in bar 6; the other caesurae in bars 2 and 4 are obvious. The same is true of the largo in Sonata Op.2 No.2.

And to the same category belongs the steady flowing eight-bar theme of the allegretto in Sonata Op.10 No.2 (Ex.120b). The repetitions in the sketch (Ex120a) involve more obvious caesurae; here again we have a development to the eight-bar C-type.

In the first movement of the Moonlight Sonata (Op.27 No.2) and in the allegro of Op.28 the caesurae are obliterated by the regular rhythm used in the accompaniment.

The theme of Sonata Op.31 No.3 is instructive. Here the melody is full of caesurae, to which the seventh-chords, with their forward urge, form a marked contrast.

As further examples may be mentioned the theme of the first allegro in Op.74; the minuet from Op.59. No.3 (Ex.88), which in spite of being constructed on the S-type shows scarcely any caesurae; the third movement of Op.18 No.3; and the slow movement of Op. 18. No.1, with its accompaniment in triplets. I have purposely chosen early examples: in these Beethoven does not so completely avoid caesurae as he does in the later works already mentioned.

Finally, there is still another method of eliminating the caesurae - i.e. by means of polyphony. Thus in bar 3 of the cavatina from Op.130 the accompaniment carries on the feeling of movement during the caesura in the violin theme.

In the same way the interpolation marked + at the beginning of Op.135 (Ex.124) bridges polyphonically the halts and rests in the theme.

(h) The fugue theme.

Considered from this point of view Beethoven's tendency to use fugato and fugue in his later works takes on a fresh aspect; for nothing is so well adapted to eliminate caesurae and pauses as is fugato, where the new voice can come in just when a caesura would occur. Even in the early works we find experiments of the kind here and there; as, for instance, in the scherzo of Sonata Op.2 No.3 (Ex.125).

Or, again, in the presto of Sonata Op.10 No.2 (Ex.126).

In considering these movements we see something new, which, however, is on the same lines as the fact of a regular number of bars in the C-type melodies, already referred to. The length of the fugue theme is such that a phrase of 8, 16, or 12 bars results. We can think of the theme from Op.2 No.3 (Ex.125) as evolved from a threefold repetition on the plan 3x2 + 2 = 8; similarly, the presto theme (Op.126) has the structure 3x1 + 1 = 4 in triplicate. The fugue and fugato themes of the later period also show the same regularity in the number of bars. That from Op.131 (Ex.10) is in four bars, and is repeated four times in succession.

The fugal working-out in the last movement of Sonata Op.101 (Ex.127) shows an eight-bar theme, the final shape of which depended on sequence-contraction, as is clear from a sketch given by Schenker.

The theme of the fugue in Sonata Op.110 is in four bars, with the structure 3x1 + 1; an irregularity comes in with the entry of the third part, since the second part lengthens the theme by two bars. The theme of the second movement in the ninth symphony is also in four bars, and it is easy to find other examples in the Missa solemnis and the fugue for string quartet Op.133.

Of course, we find instances of different structure, such as the irregularity in the entry of the parts in Op.110, to which I have just referred. But in comparison with the variety in bar-number and number of entries of the fugue themes of Bach's 'Well-tempered Chavichord' we are justified in making some such statement as the following. In Beethoven the theme of the fugue or fugato, and also the passage that involves the entry of the several parts (exposition) usually display the same regular number of bars as do themes of the S- and C-types.

This concludes what I have to say regarding the study of thematic form yielded by a comparison of the sketch with the completed work. Our investigation has shown that in many cases where the motive is retained a great portion of the work is expended on the production of form; our study has yielded us the types and the laws governing these forms, and we have learnt more about them than has been known hitherto.

Our study, however, has been based on the works of Beethoven alone; so that for the time being the points we have discovered concerning style must hold good for him alone. We must reserve for a more extensive study the detailed demonstration of the differences between, say, the classical composers of the Viennese school; in the foregoing I have hinted here and there at the lines that this investigation should follow. I believe that, especially as concerns the C-type, it will be found that similar differences occur in the number of bars used for the theme and in the formation of mixed types, as I have already shown they do in the case of the melody based on triplication (p.46 et seq.). I find support for my view in Essner's observation that 'continuation-types in eight bars are rare among the works of Haydn'.