Quote Originally Posted by AgX View Post
But that should only deliver right results at the calibration temperatures (test object temperature and ambient temperature).

For different situations see here:

http://translate.google.de/translate...Fadenkorrektur (automated translation from German)
That is correct, but the fundamental "stem correction" may cause variations at the outset and this is often why people see differences in thermometers.

Another problem is the use of alcohol and mercury in thermometers. Even though the markings on the thermometer may have been adjusted for the different liquids, the use of different liquids (and dyes in the alcohol) may affect the coefficient of expansion.