

      
Rounding Procedure Presentation of the "Rounding Programme"
The following Excel file contains a computer programme in order to round the results of germination test. The procedure included in this program is in accordance with the prescriptions of ISTA Rules (Chapter 5, 5.8.A) Rounding Programme
People involved in the project: JeanClaude Lalande, Sylvain Grégoire, Joël Léchappé, Sylvie Ducournau, Valérie Blouin, Philippe Garreau, Pierre Soufflet (France) Germination Rounding Procedure (article published in Seed Testing International No. 130, October 2005)
Calculation of Water Retention of Germination Media Presentation of the "Water Retention Calculation Programme"
The following Excel file contains a computer programme in order to calculate the water retention of germination media. The procedure included in this programme is in accordance with the guidance given in the ISTA Handbook on Seedling Evaluation (Appendix 5: A5.3 Germination Procedures  Growing Media Specification Checks  Water Retention).
Water Retention Calculation Programme
People involved in the project: Ronald Don, Sylvie Ducournau, Joël Léchappé, Gillian McLaren
Tolerance and Confidence Interval Calculator for Germination Tests Presentation of the "Tolerance and Confidence Interval Calculator"
The following Excel file contains a computer programme that calculates: tolerances of germination test replicates; repeat germination tests carried out in one laboratory; and repeat germination tests carried out in different laboratories. Confidence intervals for germination test results can also be calculated using this tool. The calculations carried out by this programme are in accordance with those developed and published by Miles 1963 and published in the “Handbook of tolerances and measures of precision for seed testing” Proceedings of the International Seed Testing Association, 28, 525686. The Excel file contains 3 worksheets:
 Germ Tol  1 lab
 Germ Tol different labs
 CI for Germ
Worksheet 1: Germ Tol 1 lab.
This should be used for the calculation of tolerances for replicates and repeat tests carried out in one laboratory. To use the worksheet enter the number of tests or replicates, the number of seed per test or replicate and the average germination which will be rounded down automatically to the nearest whole number for values below x.5 and rounded up automatically for values above or equal to x.5. For example: A. If we conduct a 400 seed germination test and obtain the following results for the four replicates: 



The average is 94.25. We enter 4 (# of tests or replicates); 100 (#of seed/test or replicate); and 94.25 (Average germination) into the calculator. The Reported germination is 94 and the Maximum range is calculated by the worksheet as 10.Since the maximum difference between replicate results is only 4 the results are in tolerance.
B. If we conduct a 400 seed germination test and obtain the following results for the four replicates of 100 seeds: 
The average is 93.5. We enter 4 (# of test or replicates s); 100 (#of seed/test or replicate); and 93.5 (Average germination) into the calculator. The Reported germination is 94 and the Maximum range is calculated by the worksheet as 10.Since the maximum difference between replicate results is 17 the results are out tolerance.
C. If we conduct two 400 seed germination tests on a sample in the same lab and obtain the following results for the 2 tests: 
The average is 96.5. We enter 2 (# of tests or replicates); 400 (#of seed/test or replicate); and 96.5 (Average germination) into the calculator. The Reported germination is 97 and the Maximum range is calculated by the worksheet as 3. Since the maximum difference between the test results is 5 the results are out tolerance.
D. If we conduct three 400 seed germination tests on a sample in the same lab and obtain the following results for the 3 tests: 



The average is 96.33. We enter 3 (# of tests or replicates); 400 (#of seed/test or replicate); and 96.33 (Average germination) into the calculator. The Reported germination is 96 and the Maximum range is calculated by the worksheet as 4. Since the maximum difference between test results is 5 the results of the tests are out of tolerance.
E. When using the tool to calculate the tolerance for 2 replicates of 50 seeds in a 100 seed test, the replicates of 50 seeds are expressed as a percentage before one uses the calculator. If you do a 100 seed test in 2 replicates of 50 and get the following result: 
To test the tolerance the results of the two reps are expressed as percentages: 
The average is 85. We enter 2 (# of tests or replicates); 50 (#of seed/test or replicate); and 85 (Average germination) into the calculator. The Reported germination is 85 and the Maximum range is calculated by the worksheet as 16. Since the maximum difference between replicate results (expressed as a percentage) is 10 the results are in tolerance.

Worksheet 2: Germ Tol  different labs.
Tables giving tolerances for tests conducted in different labs are not required in the Rules, however, for statutory purposes, one often wants to compare the results from different labs. Also when conducting comparative tests there is sometimes the requirement to assess whether the results are in tolerance. The worksheet should be used for the calculation of tolerances for repeat tests carried out in different laboratories (up to ten). To use the worksheet enter the number of tests, the number of seed per test and the average germination which will be rounded down automatically to the nearest whole number for values below x.5 and rounded up automatically for values above or equal to x.5.
For example:
A. If we conduct a 400 seed germination test in one lab and repeat the analysis in another laboratory and get the following results: 



The average is 81. 5. . We enter 2 (# of tests); 400 (#of seed/tests); and 81.5 (Average germination) into the calculator. The Reported germination is 82 and the Maximum range is calculated by the worksheet as 9. Since the difference between the lab results is 7, the results of the two labs are in tolerance.
B. If we conduct a 200 seed germination test in one lab and repeat the analysis in another laboratory and get the following results: 
The average is 88. We enter 2 (# of tests); 200 (#of seed/tests); and 88 (Average germination) into the calculator. The Reported germination is 88 and the Maximum range is calculated by the worksheet as 10 Since the difference between the lab results is 20, the results of the two labs are out of tolerance.
C. If we conduct a comparative germination test on a sample of seed and there are 8 participating labs each carrying out a 400 seed germination test and get the following results: 
The average is 91.875. We enter We enter 8 (# of tests); 400 (#of seed/tests); and 91.875 (Average germination) into the calculator. The Reported germination is 92 and the Maximum range is calculated by the worksheet as 9 Since the difference between the lab results is 10, the results of the labs are out of tolerance. 
Worksheet 3: CI for Germ
Tables giving confidence intervals for specific germination results are not required in the Rules, however, often customers ask for this information and labs find this information useful when explaining the meaning of results to their customers and the need for testing a specific number of seeds. To use the worksheet enter the number of seed tested, the level of confidence required, and the average germination, which will be rounded down automatically to the nearest whole number for values below x.5 and rounded up automatically for values above or equal to x.5.
 the 95% confidence interval is required for a germination of 80% on 100 seeds. We enter 100 (# number of seeds in sample tested); 95% (Level of confidence (%)); and 80 (Average germination) into the calculator. The Reported germination is 80 and the Confidence Interval is calculated by the worksheet as 71 – 87.
 The 99% confidence interval is required for a germination of 80% on 100 seeds. We enter 100 (# number of seeds in sample tested); 99% (Level of confidence (%)); and 80 (Average germination)into the calculator. The Reported germination is 80 and the Confidence Interval is calculated by the worksheet as 68 – 89.
 The 95% confidence interval is required for a germination of 79.5% on 400 seeds. We enter 400 (# number of seeds in sample tested); 95% (Level of confidence (%)); and 79.5 (Average germination) into the calculator. The Reported germination is 80 and the Confidence Interval is calculated by the worksheet as 76 – 84.
 The 95% confidence interval is required for a germination of 80% on 1,000 seeds. We enter 1,000 (# number of seeds in sample tested); 95% (Level of confidence (%)); and 80 (Average germination)into the calculator. The Reported germination is 80 and the Confidence Interval is calculated by the worksheet as 77 – 82.
 The 95% confidence interval is required for a germination of 94.25% on 400 seeds. We enter 400 (# number of seeds in sample tested); 95% (Level of confidence (%)); and 94.25 (Average germination) into the calculator. The Reported germination is 94 and the Confidence Interval is calculated by the worksheet as 91 – 96.
Tolerances Calculation Programme
People involved in the project: JeanLouis Laffont, Ronald Don, Germination Committee Members
Germination: Calculation and expression of results Presentation of the Calculator that can be used to determine the germination result that should be reported
The following Excel file contains a computer programme that calculates: the result that should be reported after testing the germination of a seed lot. The programme will inform the user whether a result can be reported or whether a retest is required. When retests are carried out the programme determines the result that should be reported or whether a further retest is required.
The calculations carried out by this programme are in accordance with those developed and published by Miles 1963 in the “Handbook of tolerances and measures of precision for seed testing” Proceedings of the International Seed Testing Association, 28, 525686.
Download the description of the Programme
Download the germination result calculation Programme (Excel 2007)
People involved in the project: JeanLouis Laffont, Ronald Don, Sylvie Ducournau 




