The paper deals with the applicability of the extended N2 method to base-isolated plan-asymmetric building structures. The results obtained by the usual pushover analysis of a 3D structural model are further combined with the aid of linear dynamic (spectral) analysis to account for the dynamic effects caused by structural asymmetry. In the paper, the method has been applied to the seismic analysis of a mass-eccentric four-storeyed RC frame building isolated with lead rubber bearings. Three different positions of the center of isolation system (CI) with respect to the center of mass (CM) and the center of stiffness of the superstructure (CS) were considered. The response was analyzed for three different eccentricities, three different torsional to lateral frequency ratios of the superstructure, and two ground motion intensities. The stiffness of the isolators was selected for three different protection levels, which resulted in elastic as well as moderately to excessively damaged superstructure performance levels. The results are presented in terms of the top, base and relative displacements, as well as the stiff/flexible side amplification factors. A more detailed insight into the nonlinear behavior of the superstructure is given in a form of ductility factors for the flexible and stiff side frames. The results of the extended N2 method for selected lateral load distributions are compared with the average results of nonlinear dynamic analyses. It was concluded that the extended N2 method could, with certain limitations, provide a reasonable prediction of the torsional influences in minor to moderately asymmetric base-isolated structures. Figure 6. Maximum ductility factors for the (a) flexible and (b) stiff side, for A10 and the CI = CM distribution. and relative displacements are presented in Figures 3-5, whereas the damage patterns are presented in Figures 6-8.Observing the displacements at the CM, it is evident that the softer isolators-as expected-result in larger top and base displacements and smaller relative displacements than the harder ones. It can be seen that the agreement with the results of NLDA is fairly good, and within the standard deviation boundaries until the eccentricities of the superstructure exceed approximately 15-20%, in extreme cases. For larger eccentricities, all the displacements obtained by the extended N2 method tend to be underestimated, especially in the case of the harder isolators, where much more damage occurs in the superstructure. The best results were obtained in the case of the 1st and PSC distributions, which give similar results, whereas the TRI distribution underestimates Figure 7. Maximum ductility factors for the flexible side, for A20 and the CI = CM distribution.Figure 8. Maximum ductility factors for the flexible side, for A30 and the CI = CM distribution. THE APPLICABILITY OF THE N2 METHOD 881 Figure 10. Maximum ductility factors for the models A10-A30 and the CI = CS distribution.to provide a better estimation of the NLDA results; however, the results obtained in ...