sating glass plates. In fig. 2.4:4 no compensation plate was used and in other tests plates were used 
during the printing procedure for the correction of optical radial distortion, atmospheric refraction and 
curvature of the earth. Effects of lack of flatness of the diapositive glass plates are evident. 
2.5 Tests of the Accuracy of Glass Grids 
Glass grids are often used for tests of instruments in practice, but it is seldom that the quality of the 
grid coordinates is specified or tested. To determine this basic factor in instrument tests several grids 
used regularly were measured in the Wild StK 824 stereocomparator. See section 2.1. 
Twenty-five points were measured in each grid. 
The computations follow the procedure described in reference 2.5:1. 
The results are summarized in table 2.5:1. 
The measurement were made by different operators, but mostly by P. Kaasila, MSE, who was also 
responsible for most of the computations. 
Reference. 2.5:1. Hallert, B.: Test Measurements in Comparators and Tolerances for Such Instruments. 
Photogram. Eng. March 1963. 
Table 2.5:1. Results of tests of glass grids 
Grid 
number 
Instrument 
S o 
microns 
dm v -dm„ 
X y 
io- 6 
Conf. lim. 
5 % level 
io- 6 
d3 
cc 
Conf. lim. 
5 % level 
cc 
1 
Zeiss stereocomp. 
2.7 
+ 16.3 
26.9 
-12.9 
17.2 
2 
18X18, 1939 
2.6 
+ 34.0 
26.3 
— 14.0 
16.8 
1 
Wild autogr. 
3.4 
— 4.2 
27.3 
-81.5 
17.4 
2 
A6, 1939 
3.4 
- 8.2 
27.1 
+ 25.1 
17.2 
1 
Wild autogr. 
1.7 
+ 16.1 
17.2 
+ 1.1 
10.9 
2 
A7, 1953 
1.5 
+ 3.3 
15.2 
+ 0.5 
9.5 
788 
Wild StK 824 
0.9 
- 5.4 
7.1 
- 0.2 
4.4 
789 
1962 
1.2 
-12.2 
10.1 
— 0.4 
6.5 
423 
Wild A8 
1.3 
+ 14.0 
9.8 
+ 8.7 
6.2 
424 
1.4 
- 4.6 
10.8 
+ 6.2 
6.8 
696 
Wild A8 
1.2 
+ 21.3 
12.5 
+ 3.9 
7.9 
697 
1.1 
+ 24.5 
10.9 
+ 2.7 
6.9 
822 
Wild A8 
1.9 
+ 10.3 
14.2 
+ 11.8 
9.0 
823 
2.2 
+ 6.2 
16.4 
+ 6.7 
10.4 
943 
Wild A8 
1.2 
+ 4.2 
9.0 
+ 11.9 
5.6 
958 
1.3 
+ 11.0 
9.8 
+ 31.8 
6.2 
The accuracy of the glass grid coordinates has improved considerably since 1939. Glass grids belonging 
to the A8 autographs were received in recent years. The average of the standard errors of unit weight 
of the 12 grids delivered since 1953 is 1.4 «m, According to section 2.1, the standard error of unit weight 
of the StK 824 stereocomparator is about 0.6 qm. The basic accuracy of the glass grid coordinates tested 
can therefore be described as having a standard error of unit weight of 1.3 qm. Further, it should be 
noted that in 6 tests out of 16 the expression dm K —dm y exceeds the confidence limit on the 5 per cent 
level and in 6 tests out of 16 the expression d(3 exceeds the confidence limit on the 5 per cent level. The 
coordinate residuals are normally distributed in 10 cases out of 16 according to the chi 2 -test on the 5 
per cent level.