14 BUREAU OF AMERICAN ETHNOLOGY [bulu 53 



Concerning accuracy of intonation according to the piano scale, 

 there is wide variance among singers, as well as in some instances, 

 among the several mtervals sung by the same person. The tran- 

 scriptions of these songs should be understood as indicating the tones 

 produced by the singers as nearly as it is possible to indicate them 

 in a notation which is familiar by usage and therefore convenient 

 for observation.^ A few additional signs are used and the peculiar- 

 ities which can not be expressed graphically are noted in the descrip- 

 tive analyses of the songs. Where a variation from the piano scale 

 was marked and was repeated m the several renditions of a song, 

 it is indicated by the sign + or — above the note, showing the tone 

 to have been persistently sung less than a semitone above or below 

 the note transcribed. In five records a faulty intonation at the 

 beginning of a song was corrected in the latter part (see Nos. 54, 

 129, 133, 146, 164). 



In the rendition of Indian music the writer finds tones which 

 correspond to intervals of the piano scale and occasionally, in the 

 same song, other tones whose pitch varies so constantly and by such 

 minute gradations that they have no equivalent in that scale. Tones 

 of the former class are capable of transcription in ordinary musical 

 notation; those of the latter can adequately be shown only by a 

 sound-wave chart, but, in the present work, are transcribed by the 

 notes they most nearly approximate in pitch. Minute gradation of 

 tone in Indian song has given rise to the statement that Indians 

 habitually use intervals of eighths or quarter tones. Intervals 

 smaller than a semitone are familiar to every student of Indian 

 music, but before it can s'afely be assumed that they form a fixed 

 part of a musical system it should be proved by mechanical tests 

 that they can be accurately repeated. Such proof is believed to be 

 lacking at the present time. It is the opinion of the writer that 

 these minutely graded tones are survivals of a less differentiated 

 vocal expression. In the present analysis of Indian music we observe 

 the tones on which a purely natural vocal expression crystallizes and 

 first coincides with that system of tones which has gradually devel- 

 oped in the musical history of the white race. 



In the early part of the investigations a few phonograph records were 

 made which were found to be '^ musically incoherent," the tones having 

 no clear relation to one another or to a keynote. On inquiry it was 

 always found that the men who sang these songs were not considered 

 good singers by the members of the tribe. In a repetition of the 

 song by a "good singer" the trend of the melody was the same, and 

 the intervals were such that the melody "made musical sense," con- 



1 nelmnoltz, The Sensatiom of Tone, translated by A. J. Ellis, London, 1885, pt. 3, p. 260. Translator's 

 footnote: "All these [scales] are merely tlie best representatives in European notation of the sensations 

 produced by the scales on European listeners." 



