Potato Yield Loss Prediction and Discrimination Using Preplant Population Densities ofVerticillium dahliaeandPratylenchus penetrans

Publisher Website
Google Scholar
Abstract
Reduced yields of potato, Solanum tuberosum, attributable to preplant population densities of Verticillium dahliae and Pratylenchus penetrans, pathogens involved in potato early dying disease, were quantified with linear regression models. Experimental microplot data consisted of controlled, factorial, pathogen inoculum levels in fumigated soil and resulting yields of potato cultivar Superior at two locations over a 5-6 yr period. Variation in yield relative to controls were explained best by regressions on the natural log of V. dahliae .times. P. penetrans. Population densities of pathogens at planting also were used in a discriminant analysis to predict yields above or below 90% of the control yield. A discriminant function correctly classified 86% of the yields from a 3-yr subset of data from one location. The rate of correct classification was 70% when the same discriminant function was validated on data not used in model development. Erection of a third category, "between 80-90% of the control yield", resulted in an overall 6% misclassification rate for the "below 80% relative yield" category as "above 90% relative yield", and a 23% misclassification rate for "above 90% relative yield" as "below 80% relative yield". Discriminant models intentionally minimized "below 80% relative yield" misclassification in order to promote confidence in the management of potato early dying.