Conversion of natural grasslands for agricultural uses and intensification of agricultural production has been a major cause of natural ecosystem fragmentation and biodiversity loss. Diversifying agricultural landscapes by adding semi-natural grasslands can potentially help couple agricultural production (i.e., providing food, fiber, and energy to a growing human population) with environmental stewardship, enhancing ecosystem health, and increasing biodiversity. To ensure long-term ecological benefits of buffers and to maintain them as suitable habitat for prairie-associated species, periodic disturbance is required to keep buffers in an early-successional grassland plant community. In this dissertation, I fill knowledge gaps about biodiversity of semi-natural grassland buffers within agroecosystems. I investigate influence of prairie-history on buffer forb communities and effects of disking and burning on semi-natural grassland buffer plant and butterfly communities. I also investigate suitability of using butterflies as surrogates of birds and plants on semi-natural grassland buffers. Prairie-history influenced buffer forb communities, and disking increased forb richness and abundance. Disturbance guild butterflies responded positively to disking, most likely due to increased availability of nectar-rich plants. Grassland guild butterflies were not impacted negatively by disking or burning. Responses of plants and butterflies to disking and burning varied between sampling years. Effects of disking in fall 2007 persisted for two growing seasons, but I observed little effects of disking in fall 2008. Butterflies, with the possible exception of Pearl Crescent, were unsuitable as surrogates for birds. In contrast, butterflies, including Pearl Crescent, showed suitable or marginally suitable correlations with plants. Results of my study should help agricultural producers accomplish environmental conservation objectives and provide science-based information for developing and refining USDA practice standards and policy.