Medusahead (Taeniatherum caput-medusae [L.] Nevski) is an invasive annual grass that reduces biodiversity and production of rangelands. To prevent medusahead invasion land managers need to know more about its invasion process. Specifically, they must know about 1) the timing and spatial extent of medusahead seed dispersal and 2) the establishment rates and interactions with plant communities being invaded. The timing and distance medusahead seeds dispersed from invasion fronts were measured using seed traps along 23 35-m transects. Medusahead establishment was evaluated by introducing medusahead at 1, 10, 100, 1 000, and 10 000 seeds · m⁻² at 12 sites. Most medusahead seeds dispersed less than 0.5 m from the invasion front (P < 0.01) and none were captured beyond 2 m. Medusahead seeds dispersed from the parent plants from early July to the end of October. More seeds were trapped in August than in the other months (P < 0.01). Medusahead establishment increased with higher seed introduction rates (P < 0.01). Medusahead density was negatively correlated to tall tussock perennial grass density and positively correlated to annual grass density of the preexisting plant communities (P  =  0.02). Medusahead cover was also negatively correlated with tall tussock perennial grass density (P  =  0.03). The results suggest that containment barriers around medusahead infestations would only have to be a few meters wide to be effective. This study also suggests that promoting or maintaining tall tussock perennial grass in areas at risk of invasion can reduce the establishment success of medusahead. Tall tussock perennial grass and annual grass density, in combination with soil data, may be useful in predicting susceptibility to medusahead invasion.