430 
Qa,«i = Qf'r 9'r - 0,648 
a 2 
Q(p(p Q wo, — 
5c 2 
6r' 4 
Qwjwx Qcj r a) r 0,312 
Q(pa<l'a Q(Oa(iJa 
-2 
a 2 
C 2 
r'4 
where c — 153,3 mm, r' = 150 mm, a? = 100 mm, b = 92 mm, 
b 
d — 100 mm and ~ = 0.6. 
h 
After inserting the weight coefficients into (20) and if s 4 = s r it is 
found: 
c - ' c , * „h 2 5 1 c 2 5c 2 
01_ 02 V s °Vi^V+ s - a648 ^ + S »W s v 
c 2 
'4 
S o1 — so2 = 
V 
,h 2 2 
b 2 
c- 
5c 2 
S ° b 2( 3b 2+ 4d 2 + 8d4 )+Sl2 0-312 ao 4 ^ S ° ,2 6r' 4 
+ s 0 
a r 
The above formulas give the accuracy of the inclination angles on 
the condition that the standard error of unit weight is of equal size in 
the whole photograph. It has been, however, found that the standard 
error of unit weight grows with an increasing radius [17]. The corres 
ponding weight variation has been noted in consequence of the per 
formed vertical parallax measurements at 15 points. An obvious and 
significant weight variation with the distances of the points from the 
base was noted by means of combinations in two groups of 9 points 
and computations of the standard errors of unit weight for both of the 
groups. The mean square root of all the values was 5.3 microns in the 
inner group and 8.8 microns in the outer group. P 4 and P 3 being the 
weights for y = 0 and y = ± d it is found in accordance with 
[17] and (23): 
S 01 
h 2 
1 
c- 
5c 2 
P oh 2 12d 2 b 2 
+ 19. ) + s i 2 6-648 + s 0 ' 2 —■+ 
aw 
6r' 4 
+ s c„ 
c 2 
r'4 
' S - ( “2 V S ° 2 b2 (/p 
1 (1\ + 2P 8 ) b 2 
+ )+ 
b 2 (Pi + 2P 3 ) b 2 4P 3 d 2 24PiP 3 d 4 
(24) 
i'4 
5c 2 
+ s, 2 0.312 a24 +s 0 ' 2 6r , 4 + S 9T r , 4