- 3 - 
the necessary instruments are too complicated and expensive, 
the agreement achieved in interlaboratory tests of OTF's is 
not satisfactory. 
Instead of stating a few resolution numbers which are fairly 
informative to cover all information about imaging quality, 
hundreds of OTF-curves are necessary. 
These statements ere SU partly true and to a large extent no 
longer valid today. Of course, all testing instruments which 
replace human senses like the eye by objective, automatic recording 
features are more expensive and complicated than a microscope used 
by a well-trained observer, but a well-designed instrument may 
widely surpass the exactness, repeatability and speed of the best 
observer and is therefore worth its price. 8] Furthermore, inter- 
laboratory tests have shown good agreement also for large field 
angles up to !5 degrees, and nearly all reasons of earlier dis- 
crepancies are known. It should also be emphasized that similarly 
critical and sensitive tests on optical systems in order to compare 
OTF's have never been carried out before. For example, I do not 
know an interlaboratory test of Twyman-Green out-of-axis inter- 
ferograms, although such tests might have been carried out in many 
laboratories since 1935 and are still in use today more than before. 
However, Twyman-Green tests are not restricted in their use by 
the fact that no interlaboratory tests with the same lens are being 
made. I like to emphasize that a distinction should be made between 
errors introduced by inaccuracies of the test bench, mechanical 
holders, threads, inexact definitions of the light distribution 
and the position of the image plane and such errors originating 
from the principle of OTF-measurements. According to my experience, 
the greatest differences in OTF's,measured on the same lens by 
different laboratories, are caused by experimental errors rather 
than by errors regarding the principle of OTF-measurements.