A fundamental concept of carbon dioxide influence on global temperature has been the assumption that relative humidity remains quasi-constant as global temperature rises. This supposition grew out of the Charney Report of 1979 and has been generally supported in each of the four IPCC reports and is a basic component of most of the GCM model simulations. The author questions the validity of this assumption with both observations and the theory that middle tropospheric warming results from subsidence drying. Extensive study of monthly variability of upper and middle troposhperic moisture changes associated with temperature changes from the NOAA/ECMWF reanalysis data and of the ISCCP satellite data between 1984-2004 indicates that middle and upper tropospheric absolute water vapor typically does not rise with temperature increase and that relative humidity typically goes down with temperature increases. This is consistent with the theory of upper tropospheric warming resulting from subsidence drying. A less moist upper-middle troposphere lowers the radiation emission level and leads to higher OLR and less temperature rise.