A dust insulation model for the atmosphere of Venus is proposed in which the high surface temperature results primarily from a shielding of energy escaping from the planetary interior. The insulation is provided by micron-sixed dust particles which may be kept airborne by mild turbulence. For an outflow of planetary heat of the same order as that on Earth, the required infrared opacity of the dusty atmosphere is 10 and the same atmospheric structure accounts for the observed microwave spectrum. The dust insulation model predicts a systematic variation of radar reflectivity with wavelength and the observations are in good agreement. The otherwise anomalously low value of the differential polarixation measured at 10 6 cm is expected in this model due to atmospheric absorption. The results indicate that the microwave phase effect is primarily an atmospheric phenomenon and hence the conclusions which have been drawn from it on the assumption that it is a subsurface effect are in doubt. If the cloud particle properties observed in the visual region (high particle albedo and strong anisotropy of scattering) exist throughout the atmosphere then it is possible for the incident solar energy to cause a small surface temperature variation despite the huge optical thickness of the atmosphere.