Retests shall be made as follows:
When the range of 100-seed replicates of a given test exceeds the maximum tolerated range in the table appearing in this section.
When at the time of the prescribed final count there are indications, such as presence of firm ungerminated seeds, that a satisfactory germination has not been obtained;
When there is evidence that the results may not be reliable due to improper test conditions, errors in seedling evaluation, the presence of fungi or bacteria, or inaccuracies in counting or recording results;
When a sample shows seedling injury or abnormality as a result of chemical treatment, of exposure to chemicals, or of toxicity from any source. (Retest shall be made in soil or a mixture of soil and sand);
When no two satisfactory tests are within tolerance.
To find the maximum tolerated range, compute the average percentage of all 100 seed replicates of a given test, rounding off the result to the nearest whole number. The germination is found in the first two columns of the table. When the differences between highest and lowest replicates do not exceed the corresponding values found in the “4 replicates” column, no additional testing is required. However, if the differences exceed the values in the “4 replicates” column, retesting is necessary.