Abstract : A comparison of detrital zircon ages fromthe Cenozoic forearc basins in SW Ecuador–NW Peru and those derived from theAmotapes Massif offers a singular opportunityto shed some light on the evolution ofthe NW Andean margin of South America.An extensive set of pre-Andean U-Pb zirconages (older than Late Cretaceous) wasobtained from 18 detrital samples comingfrom the Talara, Zorritos, Progreso, andGulf of Guayaquil–Tumbes forearc basins,as well as from two metasedimentary rocksof the Amotapes Massif. U-Pb age analysison detrital zircon grains allows a generalsource-to-sink scenario to be determinedfor this section of the Andes. These analysesgive consistent insights into the early topographicgrowth during the modern Andeanorogeny, the occurrence of proto-Andeanmagmatic periods in the western margin ofGondwana, and the evolution of the AmotapesMassif within the geodynamical settingof northwestern South America. Zircon agesat ca. 320 Ma and Neoproterozoic clustersfor the Amotapes metasediments show significantsimilarities along the strike of themetamorphic belt, thus defining a commonmetasedimentary origin, and can be comparedwith previously dated rocks from theEastern Cordillera of Peru. These strong similaritiesdemonstrate the presence of a wide,polyphased metamorphic belt that includesthe western parts of the Eastern Cordillera,the northern section of the Occidental Cordilleraof Peru, and the Amotapes Massif.Detrital sediments of the forearc are markedby a major gap in Early Cretaceous–Jurassicages and the minor presence of Ordovician(Famatinian orogeny) and Carboniferous–Permian (340–285 Ma) ages. The age spectrumis not compatible with sediment inputfrom the eastern Andes and suggests that theexternal areas of the Andean chain (e.g., theWestern Cordillera) were already upliftedprior to or immediately after forearc initiationduring the Paleocene, thus representinga pervasive drainage divide throughout all ofthe Cenozoic. The Amotapes Massif or otherterrains sharing the same age affinities (e.g.,the Olmosterrane; Occidental Cordillera ofPeru) are likely to represent the best potentialsources for pre-Andean zircon grains in theforearc. Age clusters in both source and sinkat ca. 600 Ma, 570–560 Ma, and 520 Ma mostlikely define magmatic periods and may representfirst-order constraints for the difficultcorrelation between proto-Andean magmaticand orogenic periods to the north and southof the Huancabamba deflection. It is unlikelythat the Neoproterozoic zircon grains camefrom the Pan African–Braziliano belts, thussuggesting the onset of significant magmaticactivity in this section of the Andes at around570–560 Ma and probably ca. 600 Ma. The520 Ma cluster is compatible with the mostprominent magmatic period in the southernAndes related to the end of the Pampeanorogeny, thus suggesting the existence of aregional proto-Andean magmatic arc duringEarly Cambrian times.