A tolerance is the limit of difference between two tests beyond which difference is considered real one, in the defined circumstances.
In seed testing the tolerance represents:
The expected variation between a fixed standard and the extreme limit of calculated variation, and
The expected normal variation between two tests conducted independently of each other. The value of the seed lot may be better or worse than the result of a valid test to the extent of tolerance.
Application
The various tolerances used in connection with the rules for seed testing are:
Comparison of two tests of the same submitted sample in the laboratory.
Comparison of two tests of the same submitted sample in different laboratories.
Comparison of two tests in the same laboratory of two different submitted samples from the same lot.
Comparison of tests in different laboratories of two different samples from the same lot (no matter which result is inferior)
Important
It is important to that the tolerances should not be confused with allowances (permissible limits) for labelling of seeds. The tolerances are provided to take care of the unavoidable variation in seed testing results and they are not to be applied prior to labelling by adding them to the results found by seed test. The tolerance should never be used for the purpose of permitting labelling to show higher quality than is actually found by the test.
Procedure of using
Calculate the average of two results to be compared. Enter the appropriate table at this value in the first or second column and the tolerance will be found opposite. If the actual difference exceeds the tolerance found in the table, the difference is to be regarded as real one.
Ag.
Technologies
(Seeds)
