INTERACTIONS BETWEEN ROOT-INFECTING FUNGI AND PLANT-DENSITY IN AN ANNUAL LEGUME, KUMMEROWIA-STIPULACEA

1 Population dynamics of an annual legume, Kummerowia stipulacea, were followed in permanent plots in a mown held for 4 years. There was a l arge reduction in the number of plants within seasons. 2 Laboratory st udies showed that biotic factors were responsible for seed death in th e soil. Using soil from the field plot location, we isolated fungi pat hogenic to K. stipulacea seeds and seedlings. 3 Two pathogenic fungi ( Rhizoctonia solani and Pythium irregulare) were used in a glasshouse s tudy of the effects of pathogens and intraspecific competition on plan t mortality and growth. Our null hypothesis was that death of individu al plants due to disease would be balanced by increased growth of surv ivors. A factorial experiment was established with five soil treatment s (three controls, P. irregulare and R. solani) and seven seed densiti es. 4 Inoculation by both R. solani and P. irregulare led to a reducti on in final numbers of plants, with the reduction greatest for populat ions inoculated with R. solani. Average biomass per plant was reduced with increased sowing density in all treatments except the R. solani i noculation, where surviving plants in high-density pots may be larger because of the reduced density resulting from pathogen-induced death. No treatment differences in size hierarchies were evident. 5 Surviving plants in both inoculation treatments could not take full advantage o f the reduced density, and thus total biomass and number of trifoliate leaves per population (a surrogate for seed production) were reduced compared to controls, leading us to reject the null hypothesis. The re duction was more severe at high sowing densities. These results sugges t that disease has the potential to reduce seed production and thus pl ant population numbers, if populations are seed limited. Shapes of yie ld-sowing density relationships were asymptotic for all treatments; th us, disease-induced 'overcompensation' leading to fluctuating populati on dynamics was not suggested.