This passage deserves close examination. It will be seen that in the voice parts we find not only a regular fugal exposition, but an almost strict canon, 4 in 1, in the fourth and octave below, the slight deviations being those necessitated by the tonal answer. At the same time the violin parts, for the sake of which the extract is quoted, have a different canon, 2 in 1, also very nearly strict, and these close imitations are carried on in as easy and natural a manner as possible. The passage is worthy of old Bach himself.
445. It will be seen that this chapter consists almost entirely of examples. It is only in this way that assistance can be given to the student; for in every branch of practical composition the invention of the melodies and the counterpoints must be left to the composer himself. By seeing what others have done before him, the student will be stimulated to go and do likewise. A few general inferences may, however, be drawn from the examples we have given.
446. First and foremost is the necessity of clearness, on which we have already so often insisted, the importance of which increases as we add to the number of parts. As an illustration of this, look at the passage by Bach in § 437, and note how the orchestral accompaniments are written in such a way as never to obscure the progression of the voice parts by which the fugue is carried on. The same thing will be observed in the fugue by Haydn quoted in § 440.
447. If the accompaniment to a fugue is anything more than a mere filling up of the harmony—that is, if it has independent features of its own, as in the examples to §§ 441, 442, 444, it is very desirable that it should be well contrasted with the subject and counterpoint of the fugue itself.
448. It is seldom advisable to have more than one, or at most two, free instrumental parts in the accompaniment to a fugue, because not only is the difficulty of composing much increased, but each added part renders it less easy to preserve clearness. It is quite true that we often find more than two free parts added in Bach's vocal fugues; but it must not be forgotten that it seems to have been about as easy to Bach to write in ten or twelve real parts as it is to the average composer to write in four. The student will do well not to overtax his strength by attempting feats too difficult for him. Unless he can write florid counterpoint fluently in at least five or six parts, he had better not try to compose an accompanied fugue at all.
449. Beyond these general hints, it is not possible for us to go. The farther we advance in this series of theoretical works, the more we are compelled to leave the student to his own resources. In our earlier volumes on Harmony and Counterpoint, it was possible, at all events in most cases, to give very definite rules as to what to do, and what not to do. These subjects bear the same relation to composition that grammar
