The Parana, one of world's largest rivers, drains two main geological domains with contrasting topographic and climatic conditions: the warmer and wetter Parana intracratonic basin in Brazil and the cooler and drier Andean fold-thrust belt and Chaco retroarc basin in Argentina. Such geomorphological differences are reflected in the mineralogy and geochemistry of clay and fine silt generated in these two contrasting domains. In tropical Brazil, kaolinite is the virtually exclusive clay mineral independently of source rocks, with smectite prevailing locally in basaltic lowlands. Extreme depletion in K, Rb, Ba and Na, very high Al, Fe and P, abundant iron oxides, and highest Ce anomaly are testified in Paranaiba mud in the north, pointing to strong lateritic weathering with complete breakdown not only of plagioclase but even of K-feldspar. In stark contrast, both plagioclase and Kfeldspar are preserved and depletion in mobile elements is minor to negligible in northern Argentina, where illite characterizes mud produced by physical erosion in the semiarid Andean Cordillera and remains dominant to as far as the Rio de la Plata. Mineralogical and geochemical parameters thus consistently testify to a steady southward gradient throughout the Parana catchment, with progressively decreasing weathering intensity towards higher latitudes. Mud mineralogy and elemental and isotope geochemistry concur to indicate that the Andes - mostly the Bermejo River - supply the vast majority of mud (but little sand) to the Rio de la Plata estuary. This implies that most sediment generated in Brazil is currently trapped in large reservoirs upstream. Integrating mineralogical and geochemical data on different grain-size fractions is revealed as an indispensable tool to investigate sediment generation in continental domains, monitor the variability of chemical weathering, and obtain key information to constrain sediment budgets.