Observed-computed residuals of Earth-Venus radar time-delay measurements from 1961 to 1966 show variations that range to over 30,000% the expected error from the best possible general relativity fit the Lincoln Lab could generate. The variations are not random but are related to relative radial velocity and intervening plasma. These variations are evidence that the relative velocity of light in space is some form of c+v and notcas predicted by Einstein's general relativity theory.