EOBEETS] ANALYSIS OF SONGS 421 



records were obtained. Only one informant gave the material, both 

 the myths and the accompanying songs. He was kept at work 

 steadily over long hom-s of concentrated effort. Under such condi- 

 tions I have frequently noticed that Indians become considerably 

 irritated, especially when they feel that they have already given the 

 desired information clearly and sufhciently. They can not under- 

 stand bemg asked to give the same performance repeatedly and 

 resent the confinement incident thereto. The songs in which these 

 accentuated curves are most appai'eut are the last of a long series of 

 repetitions, when it would be expected that the Indian was probably 

 irked. His earlier singing shows a decided tendency to smooth his 

 melodic curves; the last of a long series shows the reverse propensity. 

 These considerations, together with much experience with and long 

 observation of singing, suggest the following cjuestions: (1) Does a 

 deepenmg of the trough of do\\mward melodic cmwes mean the same 

 as sharpening the peaks of upward curves? (2) Does such over- 

 emphasis of melodic contours in singing indicate some sort of irrita- 

 tion or excitement and does a tendency to smooth their sharpness 

 come from a certain relaxation or indifference? It is not impossible 

 that these questions may be truly answered in the affirmative. At 

 least they suggest an interesting field for study. If they should 

 prove to have indicated the truth perhaps these tendencies in singing 

 might serv^e as another useful means of testing states of emotion. 



The principal tones of the last five scales of Group III appear to be 

 substantially the same as the others in that group and in Group I, 

 namely, the diatonic scale ranging roughly from dominant to domi- 

 nant, the songs beginning on the highest tone and ending on the 

 lowest, but giving most prominence in other respects to the tone thafc 

 stands in the relation of a tonic to them. The sole reason for the 

 existence of Group III as a category is this situation of scale tones 

 above and below the principal tone, and the fact that the songs 

 which tliis group represents begin on the liighest note and end on 

 the lowest. After all, the scales of Group III are those of only two 

 distinct songs, and other considerations than these lead one to 

 classify them with the scales of Nos. 1 and 7 in Group I. It is 

 worth noting that most of the pitch fluctuations from the diatonic 

 norms in the scales of Group III lie between the doininant and the 

 mediant abov^e the chief tone or tonic, just as in Group I, indicating 

 that at least with this singer this is the most unstable region of tho 

 scale. 



The scales of Songs 8, 6, and 5 and their other renditions 27, 28; 

 22, 23; and 21 form Group II, which I have placed intermediate to 

 Groups I and III merely because their range and the position of their 

 tones in relation to the principal tone are about the same as those 

 of Group I. In all the songs of Group II except one rendition, 



