Mean annual soil CO2 fluxes from normally bare mineral soil in the Amargosa Desert in southern Nevada, United States, measured with clear and opaque soil CO2-flux chambers (autochambers) were small - <5 millimoles per square meter per day - during both El Nino 1998 and La Nina 1999. The 1998 opaque-chamber flux exceeded 1999 opaque-chamber flux by an order of magnitude, whereas the 1998 clear-chamber flux exceeded 1999 clear-chamber flux by less than a factor of two. These data suggest that above-normal soil moisture stimulated increased metabolic activity, but that much of the extra CO2 produced was recaptured by plants. Fluxes from warm moist soil were the largest sustained fluxes measured, and their hourly pattern is consistent with enhanced soil metabolic activity at some depth in the soil and photosynthetic uptake of a substantial portion of the CO2 released. Flux from cool moist soil was smaller than flux from warm moist soil. Flux from hot dry soil was intermediate between warm-moist and cool-moist fluxes, and clear-chamber flux was more than double the opaque-chamber flux, apparently due to a chamber artifact stemming from a thermally controlled CO2 reservoir near the soil surface. There was no demonstrable metabolic contribution to the very small flux from cool dry soil, which was dominated by diffusive up-flux of CO2 from the water table and temperature-controlled CO2-reservoir up- and down-fluxes. These flux patterns suggest that transfer of CO2 across the land surface is a complex process that is difficult to accurately measure.