Analyses of pospiviroids in commercial seed lots of tomato and capsicum, determined by testing of 12,000 to 40,000 seeds per lot, have enabled the development of empirically-derived distribution curves for the observed prevalences of viroids in those commodities. Those distribution curves can be considered in conjunction with statistically-based estimates of detection that would be achieved using other sample sizes. Statistical calculations using binomial distributions show that sample sizes of 3000 and 9400 seeds allow detection of viroid prevalences as low as 0.1% and 0.032%, respectively, with 95% confidence. Applying those calculations to observed viroid prevalences in contaminated tomato seed lots, it is estimated that the use of sample sizes of 3000 and 9400 seeds would detect 15% and 42%, respectively, of the contaminated seed lots identified using the larger sample sizes of approximately 20,000 seeds reported in this study. It is concluded that the higher costs associated with testing of larger sample sizes represent a worthwhile investment in agricultural biosecurity.
Keywords: Pospiviroid; biosecurity; seed-borne pathogens.