-— 0A. c 
combination of them may provide an expedient overall testing strategy. As the 
various instrumental error sources exhibit different stabilities with time, the 
optimum frequencies of the relevant laboratory- and routine tests differ accor- 
dingly. These frequences can be mutually adapted for an optimum solution over 
a longer period of time, Since the routine tests are simplified versions of the 
corresponding laboratory tests, the conduction of such mixed tests is rather 
Straight forward. 
The proposed partial tests have a high power of detection and the effort in- 
volved approaches a minimums A further reduction of the effort would consider— 
ably diminish the detection power of the tests. Since all the significant error 
sources have been considered, the conducted tests appear to be sufficiently 
thorough. The simple testing procedures also provide a good repeatability of 
the testse According to the basic principles of conduction, the tests can be 
extended arbitrarily to other error sources and the power of detection can be 
increased,if so desired, by an extension of the input structures and of the e we 
corresponding samples of observations, 
Some of the detected errors can be verified by means of observations performed 
for other partial tests. An example is observations for the tests on straight- 
ness of ways, scale errors and the non-perpendicularitye Hence, this possibili- 
ty for mutual verification of errors strengthens the reliability of the tests,