Abstract The negative impact of invasive plant species on biodiversity and ecosystem functions, such as productivity and nutrient cycling has been deemed a global epidemic. To address this worldwide concern, information is needed on how the invasion process happens and how to control an existing invasion. The main aim of the research presented in this thesis was to develop a better understanding of the interacting role different mechanisms play in facilitating invasion and then link this understanding to the design of more effective control strategies. This aim is significant because traditional weed control strategies are not working. The estimated cost of controlling weeds in Australia is $1.4 billion per year in agricultural landscapes. Despite this substantial investment, invasive weed species are estimated to continue to cost the agricultural industry $2.2 billion per year in loss of yield. Current control strategies tend to focus on killing or removing an invasive plant species directly with the application of herbicides and/or mechanical removal. These strategies have proven ineffectual because the plant communities that assemble after management often remain dominated by the same invader or another. In this thesis, I use a combination of empirical and modelling techniques to investigate how disturbance regimes and competitive interactions between invasive plants and native plants can be manipulated to improve the efficacy of restoration efforts. To do this, I use the model scenario of the invasion of Eragrostis curvula (African lovegrass), an invasive grass species introduced into Australia in the early 1900s from South Africa. This species has now spread into every Australian state and territory (chapter 2). I specifically focus on two mechanisms: (1) disturbance, i.e. cattle grazing, and (2) competitive interactions. In chapter 3, I examine connections between dominance and competitive differences among African lovegrass and several functionally similar native grass species in a pasture community. To test the displacement hypothesis, I used a glasshouse competition trial to investigate interactions between African lovegrass and two non-persisting native grass species (Themeda australis and Bothriochloa decipiens) with manipulations of resources, neighbour density, and establishment order. To test the partitioning hypothesis, I compared in situ water use patterns among African lovegrass and two coexisting native grass species (Aristida calycina and Aristida personata) based on the assumption that water is the most limiting resource in this system. The key finding of this chapter is that competition can have important, but contingent, impacts on dominance. Competitive differences appear to partially contribute to abundance patterns after establishment, but may be relatively unimportant during the establishment phase where disturbance appears more critical. In chapter 4, I provide evidence that the identification of mechanisms that led to an invasion, while crucial for the development of effective preventative measures and understanding the invasion process, may not be necessary for the design of more effective control strategies. To examine the effects of different control strategies on African lovegrass and the resultant community, I established a large factorial field-trial with a split-plot design. I manipulated grazing, soil nutrient levels and the presence of the invader. The most common control strategy (removing the causal disturbance and killing the invasive grass), based implicitly on traditional equilibrium models, was not an effective option for restoring a desirable native community. Instead, this strategy led to the dominance of a secondary invader. The most effective control strategy was based on alternative stable states models and involved maintaining grazing, and increasing the palatability of the invader with fertilizers. The key finding of this chapter is that novel approaches for control, which consider the dynamics of the invader-dominated system, are needed. In chapter 5, I investigate the benefits of explicitly incorporating actions that manipulate disturbance (natural or imposed) into control efforts. To do this, I first developed a process model that described the dynamics of an invader whose establishment is preferentially favoured by disturbance. I then couched this model in a decision theory framework, a stochastic dynamic program, and applied a case-study of another invasive plant species, Mimosa pigra (a perennial legume shrub and pan-tropical weed). The key finding of this chapter is that strategies should not only focus on existing invader-dominated sites, but should also protect sites occupied by native species from disturbances that facilitate invasion. The research discussed in this thesis makes three key contributions to a better understanding of the invasion process and the design of more effective control strategies: 1) the search for one key mechanism is not sufficient because multiple mechanisms can interact or shift in importance to facilitate different stages of invasion, 2) a novel approach is needed to restore a more desirable native community because the dynamics of the invader-dominated system can differ from the historical native community, and 3) control efforts should be broadened in focus to include protection of the integrity of native communities from disturbances that facilitate invasion.