Modifications to the theory of a new technique for making viscosity measurements on multicomponent gas mixtures are described. The new technique involves slightly altering the composition of a gas mixture flowing through a capillary tube by adding a small stream of perturbation gas. The perturbation gas is usually one of the components in the mixture and is consequently of known composition. The pressure at the inlet of the capillary tube rises when the perturbation gas is added and this pressure increase is proportional to the flowrate increase. A short time later, the pressure changes again when the composition of the gas flowing through the tube changes. This second pressure change is proportional to the viscosity change and can be an increase or a decrease. For infinitesimally small perturbation flow rates, the ratio of the second pressure step to the first is proportional to dln mu /dX(1) where dX(i) is the change in the mole fraction of component i and mu is the initial viscosity, and the ratio is independent of whether the perturbation stream is added or removed. However, when small finite perturbations are made,there are systematic differences between the ratio of the two steps of pressure depending on whether the perturbation gas is added or removed. These differences are analysed theoretically and demonstrated experimentally using the argon-nitrogen system at 24 degreesC and at an absolute pressure of 1.32 bar.