174 
The more or less complete compliance with this exigen 
will afford a criterium of the reliance which may he placed on 
the results. In practice this will generally not be the case 
because the ten amplitudes will show differences inter se, which 
will be the greater, the smaller the real value of this quaii 
and the greater the variability of the element. : 
In calculating the constants of Bessex’s formula the am 
error is committed to which attention was called in discussing 
the Hornstein method, but it is easily conceivable that the 
expressions : | 
2578 . Ox 
Фе С ње Sik r 
6 2 
25.8 Ir 
2 (UN ese NE Cos — 
are so nearly equal to 15 and 
„ 25.8 Ir 
5 sn — па, cos 7 
26 
SCH 25.8 n£ . Sin an 
C08. ——- Др 1 — x 
26 26 
an 
D » . и 
lo zero, in consequence of n being very nearly equal to 55$ 
that the error of the calculated quantities will have no sensible 
influence. S 
Bij equalizing the eleven amplitudes thus calcuted to de 
corresponding theoretical amplitudes of formula (4), eleven gf 
tions are obtained from which the value of n has to be date 
by means of the method of the deast squares. : 
By putting | 
2n M 
1 
25.8 
the equation (4) assumes a simpler form because the 2 
ular quantities larger than 560° are done away with 
Sin 5875 sin 10887 540.8 ра), 
in (nz+ (+ 299% 12.954 4.524) 
sin 12 9Е 5 sin (587: Е fe 0.8 pa) 
