We inverted the Pn travel times to characterize the geometry of the Moho along a profile across the Central Andes (20¿ S) where previous workers have estimated the crustal thickness using receiver functions. Contrary to receiver functions, this technique is not sensitive to the crustal Vs. Therefore, the comparison of the two approaches provides valuable complementary information. Overall, our results are in good agreement with those based on receiver functions. However, some important discrepancies are observed beneath the Western Cordillera and the Subandes, where we find crusts 10-km thinner than in previous models. We confirm that the central part of the orogen appears to be isostatically compensated by the presence of a thick crust. However, at both edges, the topography probably requires additional support, low-density mantle beneath the Western Cordillera and a strong flexural support of the Brazilian shield beneath the Subandes. ¿ 2001 American Geophysical Union