1131 



A = 



E .llhkh 



Furtlier we have 



therefore 



6Wz 



1 -2vi'^ . 

 L 1 



nff.vj = LTIöv,, 





/Z(^.V=: 



:«:■ 



:7r 



= /—.(? VI 1 y.nö.vjX / '-^ .e .=+1 Hö.n. 



111 order to obtain the total probability Tl' we must integrate over 

 .r5_|_i , . . . Xry from — go to -f- oo, and over .i',, .v..^, . . . x^ resp. from 

 %^,%.„ ...%c to §1 + dJi, Ï, 4- d5,, . . . I. + ^^r ; i- ®- ^^^^ integration over 

 x^,x.^,...Xp consists in this, that in the integrant a\, ^i'.,, ... .<■; are 

 replaced by §i, 1.^, . . . $;,, while éx^, ex.,, . . . ox- are resp. replaced by 

 ég^, (f5„ . . . dgp. 



So we find, 



W = 



1 



— 2tbkkXk^ J 

 e r+i . nclvh 



00 — - 00 



.^7/(/i=:,o-|-l,...^^) 



' P 2 ** 



Jtp 1 



=1/; 





We have already calculated the coefiicients hjjc, that is to say, 

 expressed them in terms of the coefficients of the given equations 

 of substitution. Their determinant E is consecpiently also known. 

 For this latter, however, a simpler expression may be deduced. In 

 order to tiiul it we start from tiie relation 



E 



nhu tx'nbkk 



0+1 H-1 



