Measurements of soil surface CO2 fluxes are reported for three sites within the First International Satellite Land Surface Climatology Project (ISLSCP) Field Experiment (FIFE) area, and simple empirical equations are fit to the data to provide predictions of soil fluxes from environmental observations. A prototype soil chamber, used to make the flux measurements, is described and tested by comparing CO2 flux measurements to a 40-L chamber, a 1-m3 chamber, and eddy correlation. Results suggest that flux measurements with the prototype chamber are consistent with measurements by other methods to within about 20%. A simple empirical equation based on 10-cm soil temperature, 0- to 10-cm soil volumetric water content, and leaf area index predicts the soil surface CO2 flux with a root-mean-square (rms) error of 1.2 μmol m-2 s-1 for all three sites. Further evidence supports using this equation to evaluate soil surface CO2 during the 1987 FIFE experiment. The soil surface CO2 fluxes when averaged over 24 hours are comparable to daily gross canopy photosynthetic rates. For 6 days of data the net daily accumulation of carbon is about 0.6 g CO2 m-2 d-1; this is only a few percent of the daily gross accumulation of carbon by photosynthesis. As the soil became drier in 1989, the net accumulation of carbon by the prairie increased, suggesting that the soil flux is more sensitive to temperature and drought than the photosynthetic fluxes. ¿American Geophysical Union 1992