K. Pkarson 23 



(4) that in all cases much rleponds on the method by which the successivc 

 momcnts are calculated, and thafc fn-quency observations and physical measiiro- 

 nicnts reqiiire to sonie extent dirt'crent processes ; 



(5) that with a good nioment fornuda the method of nionients will give 

 for practica! purposes as good fits as the method of least Squares when both are 

 applicable ; 



(G) that für fitting parabulas the method of inoments has theoretieally the 

 same basis as that of least Squares, but in application it is much easier; 



and lastly : 



(7) that far more really depends for goodness of fit on the eqnatioii to the 

 curve selected than on the nnmber of its constants, and that the multiplication 

 of constants to improve fit is not only theoretieally undesirable, but does not 

 necessarily lead to the required residt. 



