ADDED SOIL-NITROGEN DOES NOT ALLOW RUMEX OBTUSIFOLIUS TO ESCAPE THE EFFECTS OF INSECT-FUNGUS INTERACTIONS

1. The rust fungus Uromyces rumicis and the chrysomelid beetle Gastrop hysa viridula have been considered as biocontrol agents against Rumex obtusifolius. As this weed grows in widely varying soil nitrogen condi tions, and the effect of insects and fungi on their hosts can vary gre atly with changes in plant nitrogen fertilization, we investigated the influence of three nitrogen fertilization regimes (0, 200, 400 kg ha( -1) year(-1) added nitrogen) on the efficacy of these biocontrol agent s, both individually and when combined, against R. obtusifolius seedli ngs, and on the nature of the insect-fungus interaction, Non-destructi ve measurements were made at monthly intervals from May, and plants we re harvested at the end of October. 2. The addition of 200 kg ha(-1) y ear(-1) nitrogen increased leaf area by 70% by September. At harvest, leaf weight was increased by 19% and root weight by 43% compared with unfertilized plants. There was no significant difference between the e ffect of 200 and 400 kg ha(-1) year(-1) added nitrogen, 3. Herbivory a lone caused a greater reduction in leaf area and decrease in root and shoot weight than rust infection alone. 4. The effects of nitrogen fer tilization, herbivory and infection were additive throughout the exper iment. The combination of herbivory and infection had an increasing ef fect on leaf area as the year progressed, from a 46% reduction in July to 62% in October, compared with healthy plants. Unfertilized plants with combined herbivory and infection had a 65% reduction in leaf area and weight, and a 72% reduction in total harvested weight, compared t o healthy plants with 400 kg ha(-1) year(-1) added nitrogen. 5. This i s the first experiment to investigate the effect of nitrogen fertiliza tion on such a tripartite interaction in which the response of the pla nt is measured. Increasing nitrogen fertilization did not allow R. obt usifolius to escape the effects of the insect and fungus and did not a lter the additive nature of the insect-fungus interaction. However, th e simple model developed in this paper indicates circumstances in whic h nitrogen fertilization may enable plants to escape the effects of th ese interactions, and this is discussed in relation to weed biocontrol strategies.