A CLASSIFICATION OF GROUPS 289 



must correspond in some way and somehow to actualities of 

 function. Consider figure 1: 



A close examination of figure 1, reading it from right to left, 

 i.e., in the order in which it was made, will reveal 21 beats of 

 varying amplitudes followed by the 22d beat whose amplitude 

 greatly exceeds those of its immediate neighbors. These first 22 

 beats may be said to constitute a 22 group. The 23d and 24th 

 beats are relatively small and are followed by 22 beats of greater 

 amplitude, the 22d, being again clearly larger than the 21 preced- 

 ing beats. Clearly a 2 group and two 22 groups are outlined in 

 the 46 beats which together constitute the 46 group. The 



FIG. 1 



technique of this delineation is by the method of " amplitude 

 variation," but no particular significance would have attached 

 to such variations did they not so closely correspond to predic- 

 tion, or had they occurred but once. Both the latter facts in- 

 dicate that the variations are not accidental, and this being so, 

 there must exist some logical grounds for their occurrence, which 

 cannot be other than the grounds contained in the prediction. 



In his work Swindle noted and commented on the fact that the 

 final beat of a group was often accented, that is, larger than the 

 others. He referred to the fact by the term final accent. He 

 sought to identify the accentuation with the occurrence of the 



