Abstract
Sutherland's empirical formula for the viscosity of a binary gas mixture is compared with the Chapman—Enskog theoretical first approximation. When the molecular weights of the constituents are nearly equal, Sutherland's formula is found to be a very good approximation to the theoretical result; when the molecular weights differ by a moderate amount, a formula of Sutherland's type may be theoretically impossible, but may nevertheless provide a satisfactory approximation; when the molecular weights are very different, a formula of Sutherland's type is theoretically possible, but the relative error is greater than when the molecular weights are nearly equal. Similar considerations apply to Wassiljewa's formula for the thermal conductivity. Generalizations of Sutherland's and Wassiljewa's formulas to mixtures of more than two components are not compatible with the theoretical first approximations.