In each of measures 2 and 6 of verses 1, 2, and 3; and in measure 6 of verse 4, is shown a group of three notes with an asterisk above. These groups, as shown in the notation, are B, A, G; but in measure 2 of verse 4, the corresponding group is C, B, A. In those measures marked *, the singers are very plainly striving to reach the tones C, B, A. There is that quality of tension in the voices with the accompanying forcing of tone which is peculiar to untrained singers striving for a tone near the limit of their highest range. As the tones actually sounded are neither B, A, G, nor C, B, A, but are instead a sort of compromise between the two, it is quite evident that the succession intended in each of the seven measures is the same as in the eighth or odd one, viz. C, B, A. If we assume this to be the case, it eliminates seven of the foreign G naturals shown in the notation. If, however, this conjecture is wrong, and the performers really feel that the groups in question all start on B, then the G naturals are eliminated by the glissandos. The only other G-natural is shown in measure 7 of verse 4. By comparing this measure with the corresponding measure in each of the other three verses, it will be seen that the singers have taken great pains in those verses to avoid this note which does not belong to the pentatonic scale which they are using,--evidence that they do not sense the tone in the fourth verse, where it is taken glissando. The D-flat, also foreign to the scale, occurs but once. It is in measure 3 of the top line. The glissando here eliminates this tone also, but, by comparing this measure with the corresponding measure of each of the other verses, we find the same avoidance as in the case of the G-natural,--evidence that the performers do not sense this other foreign tone. The song is therefore very markedly pentatonic in character. The assumption that the seven groups marked with asterisks do not represent the real intent of the singers, is based entirely on the "stress" heard in the record. This "stress" cannot be represented in notation. Relying on the notation alone, one would be warranted in drawing a contrary conclusion and assuming that the odd measure should be made to conform to the other seven and all read, B, A, G; or, from the phonographic record, one might assume that the compromise, previously mentioned, was the intonation really intended. Primitive peoples frequently do sing and play, quite intentionally, tones out of conformity with scale tones of present-day concert music. Such tones cannot be represented by our musical notation without resort to special signs. This is not necessary in the present case, as the falling short of true intonation does not appear to be from deliberate intent on the part of the singers, but seems to be due to lack of ability.