The southern slopes of the late Cenozoic Sierra Madre del Sur Mountains, located in southern Mexico, comprise the Xolapa complex; a Mesozoic-Cenozoic magmatic arc reaching 50-100 km inland to the north and stretching 600 km from Zihuatenjo in the west to Salina Cruz in the east. There are no quantitative data to constrain either the timing and amount of Cenozoic uplift or the possible relationships between rock exhumation and plate kinematics in the region. In order to address this problem, 38 samples of undeformed and slightly deformed plutons were collected from six N-S transects along the length of the Xolapa complex (Zihuatenejo, Atoyac, Acapulco, Pinotepa N., P. Escondido, and P. Angel), and analyzed for apatite fission track (AFT) thermochronometry; in addition, 20 (U-Th)/He ages were obtained from the Acapulco and Puerto Angel transects. AFT and (U-Th)/He data do not show consistent age-elevation correlations along any of the sampled transects, but suggest slow average exhumation rates of 0.19 mm/yr since 25 Ma in the west, increasing eastward to 0.34 mm/yr since 12 Ma. These rates suggest that mountain building commenced in the region in the Miocene or, more probably, later. The proposed eastward passage of the Cocos-Caribbean-North American triple junction leading to a plate boundary change from transform (no uplift) to subduction (uplift) in the Eocene would predict an eastward propagating Eocene uplift event, which is not apparent in our data. There is little change in exhumation rates in response to postulated reorganization of the plate boundary west of the Sierra Madre del Sur (currently the Acapulco trench). The complicated age data could be the result of extended duration in the partial annealing (AFT) or retention ((U-Th)/He) zones, or may reflect cooling too shallow to be recorded by known thermochronologic systems. The data could also be influenced by the presence of unrecognized late Cenozoic faults that may exist along the analyzed transects.