or at least only in such a special way that its consideration must be left to that particular head. But as a form in itself it has been employed largely and to a degree of great importance by all the greatest masters in the department of Instrumental Music; as by Handel, Bach, Beethoven, Schubert, Mendelssohn, Schumann, and Brahms. In most cases sets of Variations are not continuous, but each Variation is detached from its fellow, making a series of little movements like the Theme, each in the same key. But this is not invariable; for on the one hand, Beethoven produced a very remarkable set of Variations on a Theme in F (op. 34), in which the key changes for each variation; and on the other hand there are many examples of Variations which are continuous, that is, run into one another consecutively, without pause, as in the last movement of Beethoven's Sonata in C minor, op. 111, and (on a smaller scale) the slow movement of Haydn's Quartet in B minor, op. 64. It is very common for sets of Variations to have a grand Coda—frequently an independent movement, such as a Fugue or free Fantasia based upon some conspicuous- figure of the Theme; as in Beethoven's Prometheus Variations, op. 35, and Schumann's Etudes Symphoniques. There can be no possible reason for tying down composers by any rigid dogmas as to key or order of succession in the construction of a work in the form of Variations. Change of key is eminently desirable, for the succession of a number of short clauses of any sort with a cadence to each, runs sufficient risk of monotony without the additional incubus of unvarying tonality. Moreover it is impossible to resist the conclusion, based on the development of the great variations in the finale of Beethoven's Sonata in C, op. 111, those in the Sonata in G (op. 14), and those on an original theme in F (op. 34), that the occasional introduction of an episode or continuation between two variations is perfectly legitimate, provided it be clearly connected with the series by its figures. For if the basis of form which underlies the Variations as a complete whole be kept in mind, it will be obvious that the system of incessant repetition, when thoroughly established, would rather gain than lose by a slight deviation, more especially if that which follows the deviation is a clearer and more obvious version of the theme than has appeared in the variations immediately preceding it.
It will be best to refer the consideration of the general construction of Symphonies, Overtures, Concertos, Sonatas, etc., to their respective heads, merely pointing out here such things as really belong to the general question.
The practice of prefacing the whole by an Introduction probably originated in a few preliminary chords to call the attention of the audience, as is typified in the single forte chord which opens Haydn's Quartet in E♭ (No. 33 in Trautwein). Many examples of more extensive and purely musical introductions are to be found in Haydn's and Mozart's works, and these not unfrequently contain a tune or figure of some importance; but they seldom have any closer connection with the movement that follows than that of being introductory, and whenever there is any modulation it is confined within very small limits, generally to a simple alternation of Tonic and Dominant. Beethoven has occasionally made very important use of the introduction, employing free modulation in some instances, and producing very beautiful tunes in it, as in the Symphony in A. The most important feature in his use of it is his practice of incorporating it with the succeeding movement; either by the use of a conspicuous figure taken from it as a motto or central idea, as in the Sonata in E♭, op. 81a; or by interrupting the course of the succeeding movement to reintroduce fragments of it, as in the Quartet in B♭, op. 130; or by making it altogether part of the movement, as in the 9th Symphony, where it has an immediate and very remarkable connection with the first subject.
The order of succession, and the relation of the keys of the different movements of which each complete work is composed, passed through various stages of change similar to those which characterised the development of the form of the several movements, and arrived at a certain consistency of principle in Mozart's time; but contrast of style and time is and has been, since the early Suites, the guiding principle in their distribution. In the Suites and early examples of instrumental music, such as some of Haydn's early Quartets, all the movements were in the same key. Later it became customary to cast at least one movement in another key, the key of the Subdominant predominating. No rigid rule can be given, except that the key of the Dominant of the principal key seems undesirable, except in works in which that key is minor; and the use of very extraneous keys should be avoided. In Sonatas prior to Beethoven the interest generally seems to centre in the earlier movements, passing to the lighter refection at the conclusion. Beethoven changed this, in view of making the whole of uniform interest and equal and coherent importance. Prior to him the movements were merely a succession of detached pieces, hitched together chiefly with consideration of their mutual contrasts under the name of Sonata or Symphony—such as is typified even in Weber's A♭ Sonata, of which the two last movements were written full two years before the two first, and in the similar history of some of Mozart's works. With Beethoven what was a whole in name must be also a whole in fact. The movements might be chapters, and distinct from one another, but still consecutive chapters, and in the same story. Helmholtz points out the scientific aspect of a connection of this kind in the Sonata in E, op. 90, of which he says, 'The first movement is an example of the peculiar depression caused by repeated "Doric" cadences, whence the second (major) movement acquires a still softer expression.' In some cases Beethoven connected the movements by such subtle devices as making disguised versions of
