Grassland ecosystems cover a substantial proportion of land area worldwide and remain highly valuable production systems which deliver a wide range of services to human society. Grassland management by humans can reduce the long-term sustainability of these ecosystems, with negative repercussions for food production and climate change adaptation and mitigation. The aim of this thesis was to contribute to the current understanding of how common agricultural practices (i.e. grazing, liming and nutrient fertilisation) might affect the functioning of grassland ecosystems. The research project specifically addressed how long-term management (> 20 years) has affected: (1) plant carbon (C), nitrogen (N), and phosphorus (P) stoichiometry (i.e. plant C:N:P tissue ratios); (2) plant-soil-microbe interactions, and (3) plant species diversity and composition. It was found that despite plant C:N:P ratios being affected by management in predictable ways, changes were not always predictable based on changes in soil nutrient availability. Agricultural liming was the most important practice affecting multiple belowground properties including root mass, root decomposition, and root infection by arbuscular mycorrhizal fungi (AMF). Chronic nutrient fertilisation caused significant reductions in plant species richness, while liming increased plant species evenness. A key finding of the study was that soil C stocks significantly increased with decreasing plant C:N ratios in nutrient-fertilised grasslands and not with increasing plant C:N ratios in unfertilised grasslands, contrary to expectations. This study demonstrates how long-term management can significantly influence plant-soil- microbial interactions and how different combinations of grazing, liming and nutrient fertilisation may offer management solutions to promote trade-offs among multiple ecosystem services.