Abstract
Plants emit several different volatile compounds. These may be important as a source of information for herbivore foraging choices. During my post doc period I have examined the emission of volatile compounds from winter dormant shoots of birch, secondary metabolites, nutrients and feeding behavior of mountain hare and field vole. The complex composition of plant chemistry, quantitatively and qualitatively, makes it difficult to draw conclusions based on common statistical methods. Neural networks are especially suitable for analyzes of complex relationships including nonlinearities. It is thus interesting to develop neural networks for discriminating variation in plant chemistry and the effect of plant chemistry on herbivores. This project is a first step towards continued studies of herbivore foraging choices with experiments, computational neural modeling and spatial network modeling in a cooperation with Hans Liljenström at Dep. Biometry and technology, SLU (see application of cooperative project to Formas).
Sweden