Although net ecosystem CO₂ exchange (NEE) rates tend to be low in most grasslands, especially native grasslands, these regions occupy a large portion of Earth's terrestrial surface and thus represent a potentially large and significant sink or source for C. Herein we report on findings from several experiments conducted in native grasslands and improved pastures in which micrometeorological or chamber methods were used to evaluate the impact of grazing practices on NEE. In three native grasslands of the Great Plains, recommended grazing practices enhanced NEE compared to non-grazed and heavily-grazed pastures, although under severe drought, grazed pastures became net sources of CO₂. Although grazing temporarily removes a portion of the photosynthetic surface, and over a long period of time can alter plant community species composition and thus affect seasonal CO₂ dynamics, the long-term effects of current recommended grazing practices on NEE in these grasslands appears to be minor. In contrast, year-to-year differences in weather appear to exert a much stronger effect on NEE in these native grasslands compared to management. In an improved pasture in Pennsylvania, the method of forage removal (grazed vs. hayed) affected NEE rates, although year-to-year differences in NEE dwarfed management effects. Timing of forage harvesting appeared to be an important factor explaining the variable year-to-year NEE dynamics in Pennsylvania. Finally, in Alberta, conversion of cropland to perennial forage indicates the system may respond as a CO₂ source the first one or two years after establishment, followed by modest net CO₂ assimilation in the third (second production) year. On balance, results from these different experiments suggest that while recommended grazing practices may sometimes encourage CO₂ uptake, the grazing effect on NEE is variable, and year-to-year variation in CO₂ flux dynamics complicates our ability to predict the net effect of these practices.