The effect of seed sampling intensity on the representativeness of the submitted sample as depending on the heterogeneity of the seed lot

Publication Type

Journal contribution

Authors

Kruse, M.

Year of publication

1997

Published in

Agribiological Research

Band/Volume

50/2

Page (from - to)

128-145

In computer simulations the effect of sampling intensity on the representativeness of the submitted sample was shown for the ISTA sampling scheme, the AOSA sampling scheme and the sampling scheme proposed by the "ISTA Bulking and Sampling Committee" (=BSC). Four special examples of heterogeneity within seed lots were designed by assuming variations of seed quality within bags and/or between bags, which were greater than random variation. In computer simulations seed lots with these structures of variation and of different size were sampled in a two step sampling procedure following the practice of seed sampling. By means of the representativeness of the submitted sample measured as the standard deviation achieved by the sampling schemes the following conclusions were drawn: To sample a great number of bags is only efficient, if the variation between replicate primary samples from one bag is clearly smaller than the variation between primary samples from different bags. The relation between the variations within and between the bags determines the differences in the efficiency of the sampling schemes. If each bag in a seed lot has to be sampled, then all bags should be sampled equally often in particular in small seed lots. The effect of an increasing variation between bags with increasing lot size was only small because in all three sampling schemes in large seed lots sampling intensity was high enough to reduce this effect. Taking into account these results a sampling scheme for bagged seed lots is proposed. Data on the variation within seed lots are necessary to be able to weigh up the efficiency of sampling with view to the troubles of sampling.