108. We have given quite enough examples to prove that the rule as to answering dominant by tonic at the commencement of a subject is by no means so "absolute" as it is declared to be by many theorists. For this there are two reasons. First there is the general principle already referred to in § 100, that tonic harmony in the subject should be replied to by dominant harmony in the answer. This is illustrated by the examples in §§ 105, 106. Besides this, the melodic form of the subject should be kept unchanged as far as possible; and it is quite evident that in many cases the great composers felt this to be of much more importance than the keeping of an old rule which was made before modern tonality was established.
109. A further proof that but little weight was attached to the necessity for a tonal answer is found in the fact that sometimes in the first exposition of a fugue the first answer will be tonal and the second real, as in the following case—
Here the first answer is tonal, the first note being shortened (§ 57); but the second answer (still forming part of the exposition, in which strictness is expected) is real. In the later entries of a subject, we continually meet with real answers where tonal have been given at first.
110. We saw in § 104 a subject which, except the first note, was in the key of the dominant, the answer being in the key of the tonic. We have also seen, in discussing real answers, how an extension of the same relation of subject and answer rendered an answer sometimes possible in the subdominant key (§ 71). A similar answer is also possible where the first note of the subject is answered tonally, as in the following examples—
