20 
V 
*2. 
' = -dy,,+ —dy TT + % dc- ds-x'd^c + ---^-da? -(l+ •"—-) edeo - — dt- -dl- 
•tr 2 C ^ n ~ K 9 ' « a 
., .45° 2 de ç /^4^.° 2 ae „ _ 
c 10 
z 
aSl1 _ 'ÆÎ'o'^ 21 ' « vx- + e- 
( 18 ) 
(g-yQ fx' 2 +y' 2 +c‘ 
S \/x' 2 +c 2 
de i2 - (-^ 
x 
\/x' 2 +c 2 
limité) de , zlag.3,. de 
se ' 22 /—-—^-32 7 
s vx" +c 
The following notations are used in (is): 
^ e l0 = c ° rrec ^i° n to the eccentricity between the projection ray and 
the tertiary axis in the y-direction in the projection centre . 
de^ = the correction to the eccentricity between the projection ray and 
the tertiary axis in the x-direction in the projection centre . 
de^Q = the correction to the eccentricity between the primary and secondary 
axis in the projection centre. 
den = the 
correction corresponding 
to 
de io 
below 
the projection centre. 
fle l2 = 
_ tt _ 
above 
_ tt _ 
de 21 = 
_ it _ 
to 
de 20 
belo?/ 
_ tt _ 
de 22 = 
_ tt _ 
above 
_ tt _ 
de 31 = 
_ tt _ 
to 
de 30 
below 
_ tt _ 
de 32 = 
_ tt _ 
above 
_ tt _ 
Hot all of these eccentricities are adjustable. Thus they can be regarded 
as errors due partly to manufacturing methods. They should not normally, 
however, be significant. The solution with these 20 elements ha.s not been 
performed but it is possible. If the standard error of unit weight from 11 
elements is veiy large it is possible that it may be the result of these 
errors. It is known from practical experience that de^ Q can sometimes appear 
in AT autographs^ but, normally, significant eccentricity errors in autographs 
occur only in exceptional cases. Thus, normally 11 elements can be considered 
to be enough. Ho thorough investigations have, however, been performed in 
connection with this problem.