Studies of the effects of cattle grazing on bees show conflicting results. I used two approaches to evaluate these effects in the rough fescue grasslands of Southern Alberta. First, using Structural Equation Models, I evaluated how grazing, landscape and flowers affected bees. I found that grazing directly affected flowers (not bees), flowers only affected bees weakly, and landscape was more important than grazing in explaining bee assemblages. Second, I analyzed mutualistic networks to evaluate the effects of grazing on the relationship between bees and flowers. Grazing affected both bees and flowers. Specialization (H2’) was lower in heavily grazed fields, suggesting stronger effects of grazing disturbance on specialists than on generalists. Overall, landscape composition seems more relevant to managing land for pollination services, but grazing is important to consider for conservation of rare and/or specialized bees and plants.