52 BUREAU OF AMEETCAIT ETHNOLOGY [bull, fii 



exclusively in analyzing these songs has already been employed, 

 the structure of the melodies being determined by the pitch of con- 

 tiguous accented tones (see Table 8, footnote). One consideration 

 which seems to justify this usage is the fact that, when differences 

 appear in the several renditions of an Indian song, these differences 

 almost without exception are in unimportant progressions between 

 unaccented tones. Since the sole purpose of these plots is to show 

 the trend of the melodies, it seems permissible to omit from the 

 representation, not only the unaccented tones occurring in the 

 melody, but also a distinction between whole tones and semitones 

 in progressions, and a distinction between double and triple time in 

 measure-lengths. It is obviously desirable that the graphic repre- 

 sentation be as simple as possible, the more detailed observation of 

 the melodies being contained in mathematical and descriptive 

 analyses, 



A plot of each Sioux melody having been made, these plots were 

 compared, and as a result it was found that there are five types 

 which may be considered the primary outlines, a majority of the 

 others being combinations or modifications of these. The simplest of 

 these types are designated as A and B (see fig. 19), the former being 

 a descending trend with no ascending intervals, and the latter show- 

 ing what might be termed a horizontal progression, followed by a 

 descent to the final tone. These types appear throughout the series 

 and seem to have no relation to the content of the song. On com- 

 paring the plots representing types C, D, and E with the titles of 

 respective songs, it was found that the plots of songs having similar 

 titles or uses resemble each other. This suggests a relation between 

 the content and the form of the song, a somewhat tentative conclu- 

 sion, which, in another form, was presented in the study of Chippewa 

 songs (see Bulletin 53, pages 50-58). 



The song selected as an example of Type A is No. 195, a song of 

 the grass dance. Comparison with the musical transcription will 

 show the progressions, the song having a compass of two octaves. 

 It will be noted that in this song there are no ascending intervals in 

 the series of accented tones. In addition to this song, the following 

 songs contain a similar outline: Nos. 31, 56, 120, 151, 152, 210, 215. 

 Reference to the list of songs will show that these are divided among 

 almost all classes of songs, a fact which indicates that the type is 

 persistent. This is confirmed by general observation of the structure 

 of Indian songs. Plots of other songs of this t3'pe are shown on 

 pages 204, 245, 283, 419. Many other songs resemble this type in 

 outline, but contain one or more ascending intervals in the accented 

 tones. 



The example selected to represent Type B is No. 68, a song of the 

 sacred stones. It will be noted that the melody progresses at first 



