Detection and attribution studies of the temperature response to anthropogenic greenhouse gases and tropospheric sulfate aerosol have relied on the assumption that the responses to each of these forcings add linearly. Using surface temperature from three ensembles of integrations of the second Hadley Centre coupled model (HadCM2) forced with observed changes in greenhouse gases alone, the direct effect of sulfate aerosol alone, and combined changes in greenhouse gases and sulfate aerosol, we test this assumption. We examine the residual, defined as the response to the combined forcings, minus the sum of the responses to the individual forcings, and compare its distribution with that of control variability. Considering both global mean changes and changes at each grid point, we find no evidence that the responses to greenhouse gases and sulfate aerosol combine nonlinearly in HadCM2.