Mixture models of soybean growth and herbivore performance in response to nitrogen-sulphur-phosphorous nutrient interactions

1. It is widely established that plant-mineral nutrition is an important de terminant of herbivore developmental performance and behavioural preference . Unfortunately, the specific effects of minerals on herbivory have been va riable and few unifying principles have emerged. Advances in this area may be impeded in part by an experimental approach that emphasises single nutri ents without regard to nutrient ratios. 2. In this study, mixture-design experiments were adapted to the study of m ineral nutrition and herbivore performance. The interactive effects of nitr ogen, sulphur, and phosphorous on the development of soybean looper (Pseudo plusia includens) and two-spotted spider mite (Tetranychus urticae) feeding on soybeans (Glycine max) were quantified by polynomial regression. 3. Although significant effects of individual minerals were measured, the a ctual responses to these nutrients depended on the proportions of the other components in the nutrient solution. For example, over a range of decreasi ng nitrogen concentration, resulting soybean looper pupal mass first declin ed then increased when replaced by a high sulphur:phosphorous blend, but ju st the opposite response was measured when replaced with high phosphorous:s ulphur ratio. 4. Moreover, responses to mineral proportions were generally nonlinear and the effects of mineral proportion were not only species-specific, but varie d for different responses within a species. 5. These studies demonstrate that understanding the role of mineral nutrien ts in host-plant quality requires that mineral proportions be considered in addition to concentration. Mixture modelling, which is largely unknown to ecologists, is a powerful new tool that could significantly advance the stu dy of the interactive effects of mixture components, such as in plant-nutri ent blends.