Seed sampling

Sampling of aspects of crop agriculture is difficult and care must to be used to draw conclusions from results of tests of samples.  Biology of the plant and varying environments affect the predictability of the sample’s test result.  Nearly everyone participating in agriculture realizes this problem within a short time of exposure, although it is not always expressed.
 
Seed producers are aware that each seed within the production field did not have exactly the same environment and that each seed can be potentially with a different parent and environmental interaction during and after the growing season. Seed producers attempt to use field and facility methods to limit potential problems that could eventually affect performance of the hybrid in grower’s fields.   After using these efforts, the next challenge is to predict the success of these efforts to have good purity and germination. 
 
Sampling of seed usually begins after ears are dried and shells.  Methods are used to take general bulk sample by some randomizing technique. This bulk is the sized that includes those that represent different portions of the ear, the rounds tending to be at both ends and the flats in the center. This essentially is allowing checking seed with differing pollination dates that could affect purity. Shapes of the kernels also potentially affect germination viability.
 
Seed sizes are submitted to purity and germination test often before final bagging procedures have begun.  Number of seed included in sub-sample to be tested varies by testing method. Effectiveness of the test in predicting the eventual seed effect on field performance is dependent upon the sampling accuracy, sample size and testing accuracy in evaluation.  
 
Even if the initial sampling of the seed lot is done with care, there remains a randomness factor with test sample size. The percentage of seed germinating in a lot, or percent of outcross plants actually in the lot determined by the test is affected by the test size.  As summarized in https://www.statisticshowto.datasciencecentral.com/probability-and-statistics/find-sample-size,
a germination or purity test of 100 seeds showing 100% has a 95% probability of actually being between 96-100% where as if the test size was 400 seeds showing 100% purity or germination, the actual has 95% probability of being 99-100%.  If test result showed 96%. on a 100 seed test, the actual has a 95% probability of actually being between 90-99% where as a 400 seed test is probably between 94 and 98%.

Sampling of seed lots and testing methods including number of seed tested affect the accuracy of predicting the actual germination and purity of a seed lot.