Rhythm is not confined to the larger units, but is applied as an artistic device in the detailed structure. In exhortative speech some tribes develop a rhythmic form by adding a strongly accented syllable to each word. The Kwakiutl use the strongly accented ending ai for this purpose and I may imitate the impression of their speech by saying, Welcome-ai! brothers-ai! at my feast-ai! Here the time for each word group ending in _ai_ is approximately equal, or the words leading up to the ending _ai_ are at least pronounced with great rapidity when they contain a longer phrase. In the recital of myths rhythmic structure is sometimes attained by the addition of meaningless syllables that transform the recital into a chant. Thus the Fox Indians will add in the recital of the Culture Hero legend, the syllables nootchee, nootchee. A. L. Kroeber and Leslie Spier tell us that the myths of Southern California are chanted. Edward Sapir has observed the Song recitative in Paiute mythology, each animal speaking according to a definite rhythm and tune to which the text is adjusted.[134] I have recorded an Eskimo tale from Cumberland Sound in which the travels of the hero are recorded in a chant with interspersed melodic phrases.[135] In wails the repetition of the formal cry of moaning at short intervals and the rapid, even pronunciation of the recital creates rhythmic structure. [134] Song Recitative in Paiute Mythology, Journal American Folk-lore, Vol. 23, pp. 455 et seq. [135] F. Boas, Bull. A. M. N. H. Vol. 15, pp. 335, 340, tune in Sixth Annual Report Bureau of Ethnology, p. 655, no. 13. Much stricter than in prose is the rhythmic structure of song. It is often assumed that regularity of musical rhythm, which is found in most primitive music, is due to the multiplicity of motor actions connected with the music, particularly to the close relation between music and dance. It is true that primitive song is often accompanied by movements of the body,—a swinging of the whole trunk, movements of head, feet, and arms; hand clapping and stamping; but it is an error to assume that for these the same synchronism prevails to which we are accustomed. With us the stamping and clapping of hands coincides with the accent of the song. Some tribes innervate so that the innervation for the articulation and for the movement of hands or feet coincide. This habit has the effect of letting the clapping, drumming, or stamping follow the accent of the song. It is also not a rare occurrence that the rhythmic pattern of body movements and of song are not homologous, but that they are interrelated in different ways or sometimes even seem to be quite independent. Negro music as well as that of Northwest America offers many examples of this kind.