Experimental design theory provides tools to estimate parameters of nematode dynamics and yield loss functions with the highest accuracy given the available budget for the experiment.
Variables in models, which take the role of predictors, are given values and a weight so that the best result is obtained. As a side result the prediction of the response will have a minimal variance. We introduce the concept of D-optimum design and consider in our models two response variables: the nematode density at the end of the season and the crop yield. Both variables depend on the same predictor, the initial nematode density. D-optimum design applies to linear regression models. For models that are nonlinear in the parameters, as is the case in this study, a linearisation at a given point in a parameter plane is used. When the initial nematode density at the experimental units is known prior to experimentation, one may optimally allocate the treatments to the units by selecting the combination that has the highest efficiency.