Agricultural practices can stimulate soil nitrous oxide (N₂O) and carbon dioxide (CO₂) efflux that may contribute to global warming. However, the effects of conservation practices, such as grass and grass-tree (agroforestry) contour buffer strips, on soil CO₂ and N₂O emissions have not been extensively studied in temperate regions on a watershed scale. The objective of this study was to assess the effects of establishment of grass-tree and grass contour buffer strips on distribution of soil C and N fractions and soil N₂O and CO₂ efflux in three adjacent agricultural watersheds in northeast Missouri in the north central region of the United States. The three watersheds were in a corn-soybean rotation and were underlain by claypan soils. These soils are characterized by the presence of an argillic horizon that occurs between 13 to 46 cm below the soil surface and restricts drainage and root growth. In 1997, the three watersheds were randomly assigned one of three management systems: 1) cropped in a corn-soybean rotation only, 2) cropped in a corn-soybean rotation interspersed with grass contour buffer strips, or 3) cropped in a corn-soybean rotation interspersed with grass-tree contour buffer strips. In general, soils collected in 2003 from the grass-tree contour buffer strips had the highest levels of soil organic C and N fractions and soil from the cropped watershed had the lowest. Soil pH, bulk density, and organic C and N fractions were not consistently influenced by landscape position. Throughout the watersheds, surface efflux measurements of soil N₂O and CO₂ were collected regularly during the 2004 corn growing season at the upper, middle, and lower backslope positions within each watershed. The highest soil N₂O efflux in all watersheds occurred within the first 40 days after application of N fertilizer. The cumulative soil N₂O production from the cropped watershed was 2.3 and 3.8 times greater than the agroforestry and grass watersheds, respectively. Across all three watersheds, the cumulative soil N₂O production was smallest in the lower backslope position. The upper backslope position produced 80% greater cumulative N₂O than the lower backslope position. Cumulative soil CO₂ production was lowest in the cropped watershed compared to the agroforestry and grass watersheds. The lower backslope position across all three watersheds produced 32 and 40% higher cumulative soil CO₂ than the upper and middle backslope positions, respectively. A 72-day incubation study was also conducted to determine the effects of soil water-filled pore space (WFPS) (40, 60, 80, and 100% WFPS) and N rate (0 and 1.39 g KNO₃ kg^-1 soil) on soil N₂O and CO₂ efflux from bulk soil collected under each management system. The agroforestry soil with added N fertilizer and 80% WFPS had the highest cumulative N₂O production. The grass and cropped soils at 80% WFPS with added N produced 32.5 mg N₂O-N kg^-1 soil and 18.3 mg N₂O-N kg^-1 soil, respectively. In the incubation study, soils with no N addition produced relatively low amounts of cumulative N₂O at the 80% WFPS. The cumulative soil CO₂ production was highest in the grass soil compared to the agroforestry and cropped soils regardless of WFPS and N rate. The highest cumulative CO₂ production for the grass soil occurred at the 60% WFPS, and was approximately 3 times greater than the agroforestry and cropped soils at the same WFPS. The results of this study with claypan soils indicate that variations in soil WFPS, temperature, vegetation, and in concentrations of soil NO₃^- and organic C and N fractions due to the presence of the vegetative contour strips were the primary factors affecting observed differences in soil CO₂ and N₂O efflux among the watersheds. Therefore, establishment of vegetative contour strips could have an effect on production of greenhouse gases in agricultural watersheds with similar soils.