Table 5. F-values. Terrain Model. 
Point No 
71 
72 
73 
81 
82 
83 
Operator No 
1 
1 
60 
1. 
33 
3 
27 
1. 
27 
1. 
77 
1.06 
2 
0 
46 
o 
40 
0 
33 
0 
61 
o 
26 
o. 22 
3 
3 
04 
7 
79 
3 
78 
4. 
25 
16. 
25 
0.93 
4 
2 
.25 
9 
00 
3 
42 
4. 
00 
0 
87 
5.98 
5 
2 
12 
3 
3 
35 
1. 
48 
1. 
3JL 
7 
6 
0 
80 
1. 
85 
1 
35 
1. 
23 
0. 
47 
0. 86 
7 
0 
.41 
0 
57 
0 
50 
1 
35 
0. 
14 
3.49 
8 
0 
66 
0 
18 
11 
52 
2. 
67 
0. 
10 
0.34 
9 
2 
83 
0. 
57 
16 
22 
2. 
53 
0. 
60 
5.21 
Twenty underlined values are not significant. 
Table 6. F-values. Grid Model. 
Instrument 
A 7 
Zeis s 
A 8 
Wild 
Operator No 1 
1.32 
1. 15 
1.82 
1.95 
2 
0. 76 
2.20 
0.20 
0.64 
3 
4. 85 
4 
o.31 
0. 12 
3.40 
2.00 
6 
0.41 
1.07 
3.83 
1.87 
7 
0.36 
0.41 
0.33 
1.45 
9 
7.44 
0.08 
0.33 
1.5 8 
Ten underlined values are not significant. 
Conclusions. 
The relation between the precision of the three observation methods 
seems to a certain degree dependent upon the operator and his indi 
vidual training and experience. Only nine operators have taken part 
therefore it is not advisable to draw any general conclusions. In this 
investigation it is found that the binocular methods is less precise than 
the two others, which can be seen in aerial stereo as well as in grid 
models. From the aerial stereo models it has been observed that the 
more experienced operators get a higher precision with the stereosco 
pic method but less experienced operators get higher precision with the 
monocular method. In grid models the stereoscopic method usually is 
the best one. 
The ratio 1: ^ 2 between the stereo and the monocular precision 
does not hold in this investigation. If the hypothesis was right 10 per 
cent of the F-values should be expected to fall outside the acceptance 
region, but in this case approximately 60 per cent fall outside and then 
the hypothesis has to be rejected. The individual fluctuations are great, 
however, e. g. for operator 1 the ratio seems to hold. 
483