Viability of buried weed seed and computerized seed detection techniques

The longevity of seeds is a primary factor in determining potential weed population problems. Burial study plots for ten weed species were established at four dryland locations in Wyoming. Seed was buried at 2.5cm and 15cm depths, in replicates of four. Nylon mesh packets containing 100 seeds of each species were buried in evenly spaced holes. Soil was replaced and sod allowed to develop over each site. Seed packets were removed from each of the four sites in 1991, 1992, 1994 and 1996. Laboratory analysis for annual evaluation and comparison with stored samples used the tetrazolium chloride viability test.; Computer software was written to standardize viability analysis and archive data over several years of a seed burial study. Software programming to decrease human error factors used a visual interface programming package and associated toolkits. Multivariant, edge detection programing code was written to use a color flatbed scanner connected to a computer, to count and determine the viability of intact seeds and embryos. A 24 hour, 1% tetrazolium chloride soak was used on the embryos and seeds tested. Seeds were first examined by a technician using a stereomicroscope. Archived computer seed counts and viability readings were later compared to human and verification counts.; The computer had significantly higher accuracy versus the human intact seed counts for two species. Human accuracy was greater for viability detection of a single species. There were no significant differences between the two methods for the remaining species tested. Burial study results showed that average seed viability declined 70% in the first year of the study, 3% in the second and 6% between the second and fourth years of the study. Only 6% of all seed was still viable by the sixth year. Field bindweed, cutleaf nightshade and spotted knapweed retained the highest viability of the weed species tested with viability greater than 34%, 20% and 3%, respectively.