How the Seed Laboratory Obtains Representative Seed Samples
We frequently stress the importance of obtaining a representative sample of your seed lot for testing. This includes sampling intensity and storage of the composite sample. You may not be aware it, but the components of your seed lot or composite sample can separate over time with movement or shipment.
For instance, lighter seeds of a particular kind will make their way towards the top of a sample bag during transit whereas the denser seeds will move towards the bottom. Similarly, seeds of larger size will also separate towards the top of the sample whereas smaller seeds will move toward the bottom.
This type of separation can easily be observed when opening a container of mixed nuts—you will likely notice several nuts of one kind located on the top! In addition, at the bottom of the container you will find the smaller, broken pieces of nuts, much of which would be classified as inert matter in seed testing.
So, what does all of this mean? Much like the importance in sampling of your seed lot, we want to ensure that the testing that takes place in the lab is also representative. To achieve this, we obtain the required amount of a working sub-sample for purity analysis through a process called mixing and dividing.
Prior to testing, each sample is thoroughly mixed by passing the entire sample through a calibrated divider three times. Following the third passage, sample reduction takes place through successive halving and combination of half-portions until the required minimum working weight for the particular kind is obtained.
For instance, after mixing, a 1000-gram wheat sample will be divided down to obtain a minimum of 500 grams for the noxious exam. From this noxious portion, a minimum of 100 grams will be divided out to conduct the purity analysis, in which the component percentages used for labeling are determined. Altogether, this process of mixing and dividing provides randomization of the sample while eliminating bias and it lays the foundation for subsequent tests.