Preliminary results from numerical experiments designed to show the seasonal and geographical distribution of the climatic changes resulting from increased atmospheric CO2 concentration are presented. These simulations were made for both doubled and quadrupled CO2 levels with an improved version of the two-level OSU atmospheric GCM. In these experiments and in a control run with normal CO2, the solar radiation incident at the top of the model atmosphere and the sea-surface temperature and sea ice were given prescribed seasonal climatological variations. In January the globally averaged tropospheric temperature is increased with respect to the control mean by 0.30¿C (0.48¿C) for doubled (quadrupled) CO2, which may be compared with an interannual January temperature variability of 0.15¿C in the control (as measured by the root-mean-square of January monthly averages in a 3-year control integration). In July, the globally averaged tropospheric temperature rises by 0.33¿C (0.60¿C) for doubled (quadrupled) CO2, with an average warming over land surfaces of 0.71¿ C (1.04¿C); these values may be compared with a root-mean-square interannual July temperature variability of only 0.03¿ C in the control. These results are clearly due to the model's differing thermal response over ocean and continent in the summer and winter seasons. In comparison with the results from models using a swamp ocean (which absorbs no heat), the present experiments with a climatological ocean represent the opposite limiting case in which the ocean is an infinite reservoir of heat. Although the present model gives a small but significant overall warming without poleward amplification, and does not indicate a significant change of the atmosphere's general circulation or hydrological cycle, these results are due to the unrealistic suppression of water vapor- and albedo-temperature feedbacks.