A measure of the variation in F{k,G) is given by 



(iL23) M^ --^ S[F(k, 9)]^ 



6 



and since 



(11.24) M^j [f(H., e)]^de = ^ [l+l S[(a^(k))^ + (bjk))2] 



^tt/2 L n=l 



the closeness of the fit for a. given N is given by 



(11-25) Rj^^ -- Mj^/M^ 



If R-i^r is one. the fit is perfect for the available data, 



Equation (11,18) can also be put in the form of equation C11--26), 



ill ,26) f(ia, e) -- ^ I 1 + ^S^ cjix) cos(2n(e -y^)) 



for 9 (u) - IT < 9 <6,„(a) , where 



, V r 2 2 ,1/2 



(11,27) cj|x) ^ [a^ (ix) + b^ (Hi)] 



and 



1 »,i b {(J.) 



(1128) YnW = 2nt^- ^ 



The values of c^^, y . and R for n equal to Ij 2, 3, 4, and 5 were 

 computed by means of the IBM 650 for each k in Table 11.6. The results 

 are given in Table 11.11., The values of c and y areplotted as a function 

 of k in figure 11.30,, 



The values of Cj and y-j^ show a fairly smooth variation with k as 

 do also the values of C2 and y2 T^^Q values of c^ , c^, and Cr are Ioa 

 and somewhat erratic^ and the values of y . y-^, y are highly variable 



228 



)W 



