A mere succession of tones, however pleasing separately, does not make a melody; for melody depends on a definite scale and on certain relations between the tones of the scale. These relations illustrate the three modes of aesthetic unity. First, there is harmony. Tones are harmonically related when they belong to the leading chords of the key. The tones of such chords, when sounded together, are consonant. Now harmony, which is an aesthetic feeling, although not identical with consonance, which is a purely sensory relation between tones, depends nevertheless upon consonance. In order to understand harmony, we must therefore first understand consonance, and, in order to do this, we must begin by describing the experience and then look for its possible causes. [Footnote: Consult the discussions in Karl Stumpf, Tonpsychologie; Carl Emil Seashore, The Psychology of Musical Talent, chap. VII.] As for the first, consonant tones, when sounded together, seem to fit one another, almost to fuse, despite the fact that the different tones are distinguishable in the whole. This fitting together, in turn, seems to depend on a resemblance or partial identity between them. For example, the most consonant tones are a note and its octave, which are, perhaps, actually identical in quality; but lesser intervals are also alike, as for example a note and its fifth, which are more readily mistaken for one another than two dissonant tones, say a note and its seventh. As for the explanation of consonance, we know that consonant tones have identical partial tones and are caused by vibration rates that stand to one another in simple ratios. Thus in a clang composed of a tone and its fifth, the first partial of the fifth is the second partial of the prime, and the vibration ratios are as two to three. The bearing of this second fact on the question of partial identity will become clear if we consider the concrete case of a tone produced by 24 vibrations per second, whose fifth would then be produced by 36 vibrations per second, and then consider the same tone and its dissonant second, the ratio of whose vibrations is 24 to 27; in the former case, there is a common part of 6 vibrations, a fourth of the total number of the first tone; in the latter, only 3, an eighth. That identity of partial tones is not a sufficient explanation of consonance--as Helmholtz thought it to be--is proved by the fact that simple tones, which have no partials, may still be consonant. Nevertheless, an identity of partials does undoubtedly contribute to the consonance of the complex tones used in our music; ultimately, however, the final reason for consonance must be sought in some underlying identity within the tones themselves, an identity that seems to be given psychologically in their resemblance, and with which physically the simplicity of their vibration ratios probably has something to do. And that in music the feeling of harmony should depend