The vacancy solution model of adsorption which uses an activity coefficient equation of the Wilson form has been improved (1) by incorporating temperature dependency into the model, and (2) by development of a functional relationship between the adsorbate-adsorbate binary interaction parameters. This reduces the number of regression parameters and improves the predictive capabilities of the model. Comparisons of this improved model with the Flory-Huggins activity coefficient form of the vacancy solution model are presented for binary and ternary systems. For a priori predictions from pure-component data, the Flory-Huggins form is recommended. If experimental binary data are available for all pairs, the form that most accurately correlates the binary data should be used to predict higher-order systems.