A weekly, year-round nitrous oxide (N2O) and methane (CH4) flux measurement program was initiated in nine sites within the Central Plains Experimental Range in the Colorado shortgrass steppe in 1990 and continued through 1994. This paper reports the observed intersite, interannual, and seasonal variation of these fluxes along with the measured variation in soil and air temperature and soil water and mineral nitrogen content. We found that wintertime fluxes contribute 20--40% of the annual N2O emissions and 15--30% of CH4 consumption at all of the measurement sites. Nitrous oxide emission maxima were frequently observed during the winter and appeared to result from denitrification when surface soils thawed. Interannual variation of N2O maximum annual mean fluxes was 2.5 times the minimum during the 4-year measurement period, while maximum annual mean CH4 uptake rates were 2.1 times the minimum annual mean uptake rates observed within sites. Generally, site mean annual flux maxima for CH4 uptake corresponded to minimum N2O fluxes and vice versa, which supports the general concept of water control of diffusion of gases in the soil and limitations of soil water content on microbial activity. We also observed that pastures that have similar use history and soil texture show similar N2O and CH4 fluxes, as well as similar seasonal and annual variations. Sandy loam soils fertilized with nitrogen 5--13 years earlier consumed 30--40% less CH4 and produced more N2O than unfertilized soils. In contrast, the N addition 13 years ago does not affect CH4 uptake but continues to increase N2O emissions in a finer-textured soil. Our long-term data also show that soil mineral N concentration is not a reliable predictor of observed changes, or lack of changes, in either N2O efflux or CH4 uptake. Finally, from our data we estimate that annual global N2O emission rates for native, temperate grasslands are about 0.16 Tg N2O-N yr-1, while CH4 consumption totals about 3.2 Tg CH4-C yr-1. ¿ American Geophysical Union 1996