To predict erosion rates throughout the Andes, we conducted a multiple regression analysis of the sediment discharge from 47 drainage basins in the Bolivian Andes and various topographic, climatologic, and geologic parameters. These mountainous basins are typically large (17-81,000 km2; mean = 11,000 km2), often have decades of measurement data on daily water and sediment discharge, and display an extraordinary range of denudation (0.01-6.9 mm/yr), runoff (16-2700 mm/ yr), and local topographic relief (700-4300 m), yet the underlying lithology (granitic plutons, metasediments, and Quaternary deposits) can be classified into a small number of homogeneous types, and anthropogenic disturbance is limited. The steep nature of the channels precludes sediment storage, and unlike previous global studies of fluvial denudation rates, based on data compilations from very large river basins (>100,000 km2), this analysis distinguishes the sediment production in mountains from sediment entrapment within adjacent sedimentary basins. Lithology and average catchment slope account for 90% of the variance in sediment yield, and yield is not significantly correlated with runoff. However, because runoff over geologic timescales orchestrates the processes of channel network incision and sediment evacuation, climate could ultimately govern basin hillslope conditions and thereby the rates of hillslope erosion. Several theoretical geomorphic models for mass wasting are tested to assess hillslope-scale sediment yield models for the study basins. When applied throughout the Amazonian Andes, such empirical models predict an annual Andean sediment flux to the lowland Amazon Basin of 2.3-3.1 Gt. Because ∼1.3 Gt/yr of sediment reach the gauged tributaries of the mainstem Amazon River, the intervening foreland basins appear to intercept about half of the total Andean sediment discharge. © 2006 by The University of Chicago. All rights reserved.