Although the impact of Canada thistle (CT) on annual crop production is relatively well established, few investigations report on this weed's impact within perennial pastures. This field study assessed herbage yield losses within eight central Alberta pastures from 1999 to 2001. Each pasture was sampled in 1999 to quantify thistle and herbage biomass within 25 permanent plots. CT was controlled in 2000 and the response of vegetation measured in 2000 and 2001. Before removal, significant negative relationships (P < 0.05) between thistle abundance and herbage were noted at six sites. After thistle removal, herbage at several sites displayed positive responses. Both thistle density and biomass adequately predicted herbage yield loss. Yield losses due to CT can be substantial, peaking at 2 kg/ha for each kilogram of standing thistle biomass and 4.3 kg/ha with each additional thistle stem per square meter. Demonstrated yield losses were variable among sites however, likely due to factors such as heterogeneity in soils, available moisture, and variation in disturbance history or pasture vegetation composition. CT management in perennial pastures of western Canada may enhance pasture production, but further research is required to reliably predict the ability of pastures to respond.