Analysis of the coexistence mechanisms for grasses and legumes in grazing systems

1 It is widely assumed that grass-legume associations offer a way to sustainable, low input land use, with reduced environmental impact. However, a combination of both ecological and physiological principles may be needed to understand the sustainability of species balances. 2 To increase understanding of grass-legume dynamics, we developed a model that extends a recently proposed pasture model (Thornley, Bergelson & Parsons: Annals of Botany 1995, 75, 79-94) by including selective grazing and spatial considerations. Population oscillations were shown to stem from the way grasses can exploit leguminous N fixation. If the legume is a relatively good competitor for light, populations do not oscillate near equilibrium, but in the converse case, populations do oscillate. 3 Large amplitude oscillations can arise when there are sufficiently long time delays in the plant populations' responses to changes in the competitive environment. In the present model, these stem from variable internal substrate pools (of C and N), which uncouple biosynthesis from resource uptake, but other time delay mechanisms are easily envisaged. 4 Urine deposits prevent the establishment of equilibrium within patches, but spatially random urine deposition stabilizes population fluctuations at the field scale. This is because perturbations to local N cycles desynchronize patches with regard to the grass-legume population cycle. 5 Differences in the soil N environment (fertilizer input, leaching rate) determine whether the species can coexist, but where coexistence is possible, species composition regulates soil mineral N. 6 Selective grazing (herbivory) does not essentially alter the grass-legume interaction, but complex foraging trade-offs lead to herbivory effects that may seem counter-intuitive. The model has important implications for attempts to control the legume content of mixed species communities.