The first collectors of the music of the various savage tribes naturally were obliged to write it down in ordinary musical notation. But savages in their primitive melodies, like certain animals in their quasi-musical cries, continually use intervals of less than a semi-tone. In writing down these primitive melodies in our notation it has been necessary to disregard these small intervals and to treat them as accidental happenings, a mere out-of-tuneness, as it were. The note written down has always been assumed to represent the tone which the primitive singer was trying unsuccessfully to produce. But instances of these variations from the tones of the orthodox chromatic scale finally became so numerous as to give rise to the belief that savages consciously made use of quarter tones in their songs. This belief has had many learned and eloquent defenders, among whom may be mentioned James A. Davies[10] and Benjamin Ives Gilman.[11] The truth of this theory is, however, very doubtful. The conscious use of the quarter tones or intervals smaller than those in use in European music would indicate a much more refined perception of tones and their relations, a much more delicate musical ear, than is possessed by civilized Europeans. And this is hardly to be reasonably expected of savages. Moreover, during recent years the writer has examined some hundreds of Indian songs as recorded by the phonograph. Many repetitions of single songs have been examined by him. As a general rule the repetitions fail to agree in length, rhythm, or accuracy of intonation. Frequently they agree only in general contour. Any single tone is liable to vary up or down at least a quarter of a tone, and in some cases the variation is as much as a full tone. Now if the Indians consciously use quarter tones in their songs, one would expect to find a regular recurrence of these small intervals at the same place in each subsequent repetition of the song. But as such is very far from being the case, one is led to conclude that while these fractional intervals do really occur, their occurrence is much more the result of accident than of conscious intention. These conclusions in regard to North American Indian music apply, we believe, to the music of all savages. The characteristics of that which is primitive are undoubted strength, directness of expression, and consequent effectiveness, but this elemental strength is coupled with crudity, inaccuracy, and an apparent lawlessness or impatience of restraint. No matter how charming, how effective, or how interesting many of these strains of primitive music may seem to us from an ethnological point of view, it is apparent that the _mind_ of man has not yet grasped and moulded this tonal material. Primitive music does not show the effect of thought. It is merely the wild and wayward expression of emotion.