DENSMORE] CHIPPEWA MUSIC 5 



especially easy for the voice, and not directly connected with the 

 concept of the song. 



Accidentals are usually given with more firmness and accuracy 

 than diatonic tones. This may be attributed to one of two causes — 

 either these tones are recognized as deviations from some definite 

 standard, or they are individual tones impressed on the mind of 

 the singers more clearly than other tones, and are therefore given 

 more accurately. Of interest in tliis connection is the fact that 

 the phonograph record shows the octavo, fifth, and twelfth sung 

 accurately by men who give the other intervals with uncertain 

 pitch. This peculiarity is found in records made by Indians whose 

 environment is ]:)rimitive and who rarely hear the white man's music. 

 The following system of signs has been adopted: If the singer gives 

 the principal intervals of the scale correctly, but makes deviation in 

 other intervals, these incorrect tones, if raised less than a semitone, 

 are marked + ; if lowered less than a semitone, they are marked — . 

 This tonal peculiarity has been tested in the following manner: Two 

 singers were asked to repeat songs recorded about seven montlis 

 before; about 20 songs were included in this test, which showed devia- 

 tion on the same tones in the second as in the first rendition. 



Rhythm 



The rhythm of a Chippewa song is as much a matter of composition 

 as the melody and often expresses the itlea of the song. The term 

 "rhythm," in this connection, refers to the succession of measures 

 of irregular lengths, as well as to the subdivision of the measure. 



The transcription of a song is divided into measures according to 

 the vocal accent. Since a secondary accent seldom occurs, the song 

 is usually divided into measures of two or three counts. In many 

 instances each of these counts is a metric unit and is so indicated by 

 the metronome mark at the opening of the song; in other instances 

 the entire measure constitutes the metric unit. In many songs 

 there is a recurring rhythmic unit composed of one to four or more 

 measures; in other songs there is no recurring rhythmic unit and in 

 many songs of this class the entire melod}^ constitutes a rhythmic 

 unit, complete and satisfactory in itself. Continued repetition of 

 such a song gives to the entire performance the effect of a homogeneous 

 whole. 



In many cases a metronome test of the phonograph record shows 

 the drum to have the same uulse or metric unit as the melody. From 

 this fact one would expect to find that the pulse of drum and voice 

 coincide at certain points, but an analysis of phonograph records and 

 observation in the fiekl tend to show that the drum and the voice 

 are independent expressions. This is indicated in the tabulated 



