12 BUREAU OF AMERICAN ETHNOLOGY [bum.. 5n 



the singer using the i)hrases in varying order, ap])arently as his fancy 

 prompted him. Tliis is noted in the analysis of No. 105, and was 

 observed especially in the love songs. No. 100 affords an example 

 of a song the entire rhythm of whicli constitutes a unit that is com- 

 plete in itself and can not be divided. Such a song would become, 

 in its repetitions, the rhythmic unit of an extended musical perform- 

 ance. 



Finally, there is observed the speed of voice and drum, as 

 indicated by a Maelzel metronome, the number representing the 

 number of beats per minute. The method of adjusting the phono- 

 graph to secure uniform speed in recording and in playing a 

 song has been already described. Table 20 shows the metric unit 

 of the voice, the indication being usually for the time of a quarter- 

 note, though in some instances a half-note, or even an entire measure- 

 length, was the only unit by which the tone-values could be deter- 

 mined. It will be noted that the largest percentages of speed occur 

 on the numbers 96 to 104 M. M., this group being a somewhat clearer 

 indication of the natural tempo of Chippewa song than the average 

 speed of the entire collection (107 M. M.), as the latter is slightly 

 aflfected by songs whose peculiar structure necessitates a very large 

 or a very small unit of measurement. The metric unit is particularly 

 slow in songs of controlled excitement (see No. 30). 



Table 21 shows the metric unit of the drum, the highest percentages 

 being between 104 and 112 and the average speed 109. Both these 

 tests show the speed of the drum to be greater than the speed of the 

 voice, though a proportion between the two is not evident. 



The comparative speed of voice and drum is further shown in 

 Table 22, the songs in which the drum is slower than the voice being 

 about half the number of those in which the metric unit is the same, 

 and less than half the number of those in which the drum is faster 

 than the voice. The independence of the vocal and instrumental 

 expressions is further shown by the fact that the tempo of the voice 

 may change but the tempo of the drum remains the same, a peculiarity 

 which is noted in the analysis of No. 168. 



There may be instances in which the metric units of A^oice and 

 drum are in the ratio of two to three, but the A\Titer does not recall 

 an instance in Chippewa music in which drum and voice coincided 

 on the first count of the measure, one showing two and the other 

 three pulses, or metric units, during the measure, although tliis 

 " two-against-three rhythm" has been found in the music of other 

 Indian tribes and among many other primitive peoples. Fillmore 

 gives an mstance of a Bala Bala (Bellabella) Indian song containing 

 a 2-4 rhythm in the voice and a 5-8 rhythm in the drum, the two 

 coinciding on the first of each measure.^ In Chippewa music, 



1 John Comfort Fillmore, Primitive Scales and Rhythms, p. 173. 



