Paleomagnetic results from 61 sites in Tertiary red beds and volcanic rocks from the Bolivian Andes allow quantification of tectonic rotations within the Altiplano domain. A total of 16 sites were also obtained in lower Miocene ignimbrites that cover the forearc of the north Chilean Andes near Arica. In the southern Altiplano (Lipez region) a local clockwise rotation of up to 38¿ is recorded in lower Miocene volcanic rocks. Farther north, near the Salar de Uyuni, there is no evidence of significant rotations. Counterclockwise rotations are observed in the Northern Altiplano. The largest counterclockwise rotation (28¿) is found in Eocene-Oligocene red beds (Tiwanaku Formation) along the eastern border of the Corque syncline. Middle Miocene sediments located within the center of the Corque basin record a counterclockwise rotation of only 10¿. About 200 km north of the Corque basin, Eocene-Oligocene sedimentary rocks near the locality of Viacha record only 10¿ counterclockwise rotation. Paleomagnetic results in lower Miocene ignimbrites from the forearc near the Arica bend do not show evidence for late Cenozoic rotation of the forearc. These new results suggest that the 20¿ to 25¿ counterclockwise rotation of the southern Peruvian forearc occurred prior to the Miocene. Anisotropy of magnetic susceptibility (AMS) data show a magnetic fabric mostly controlled by bedding, but slight AMS lineations are found almost parallel to the major regional structural trends such as fold axes. The apparent relationship between tectonic rotations, AMS lineations, and structural trends suggests that rotations occured after or during the final stage of folding. Paleomagnetic data obtained in Paleozoic rocks show evidence of clockwise rotation of the southern sub-Andean ranges. Tectonic rotations during the Neogene are mostly localized on the eastern side of the central Andes. The curvature of the occidental margin near Arica was likely acquired prior to the last stage of Andean deformation. Although the magnitude of the rotations may vary from one locality to the other, there is a consistent pattern showing counterclockwise rotations to the north and clockwise to the south. A large-scale regional mechanism is needed to explain this pattern of tectonic rotations within the central Andes. We propose a model in which along-strike segmentation of the Andean foreland and indentation of the Altiplano by a curved forearc are the major factors controlling tectonic rotations within the Altiplano, Eastern Cordillera, and sub-Andean ranges. ¿ 2000 American Geophysical Union