10 
Appendix I 
1:1. Some Examples of the Application of Tolerances in Practice 
1:1.1 Standard Error of Unit Weight 
Stereocomparator 
In Hallert 1963 some principles of tolerance determination of stereocompara 
tors have been discussed. A comprehensive test program of some stereocom 
parators was carried out. The average of the standard error of unit weight from 
a set of preliminary tests of one specific stereocomparator was found to be l.o 
micron and was temporarily used as representing the basic geometrical quality 
of the type of instrument in question. According to the F-test the standard 
error of unit weight in another instrument of the same type, tested with the 
same grid and under identical circumstances (25-points test) should not exceed 
1.2 times the value l.o micron, i.e. 1.2 micron. Measurements in 25 grid points 
by 11 different operators and adjustments according to the method of least 
squares, gave for 33 tests the following distribution of the standard errors of 
unit weight: 
Standard Errors of 
Unit Weight, micr. 0.7 0.8 0.9 l.o l.i 1.2 
Number 1 8 11 7 3 3 
In no case, consequently, was the tolerance of 1.2 micron exceeded. 
Autograph A7. 
According to the international tests as reported in Hallert-Ottoson-Ternryd 
1960 the standard error of unit weight of projector coordinates of autographs 
A7 was on an average found to be 5 microns (about 300 redundant observa 
tions). The tolerance for the standard error of unit weight as determined from 
measurements in 9 points and adjustments using 8 parameters is 1.4 times the 
value 5 microns i.e. 7.0 microns. 
In the autograph A7 No. 310 grid measurements have been made about 
once a month for checking the mechanical adjustment and for determining 
the basic accuracy. During a long period (approximately 5 years) the standard 
errors of unit weight were found to be smaller than the mentioned tolerances. 
As an example of such a test measurement and computation of the standard 
error of unit weight, the results shown in Example No. 1 in Kaasi/a 1961 can 
be mentioned. The measurements were made 21 April 1961 and the standard 
error of unit weight of the projector coordinates was found to be 4.6 microns, 
but during the latter part of 1962 the standard errors of unit weight in the 
right projector were found to be 9.7 and 8.8 microns respectively, indicating 
a decrease of the basic accuracy.