Abstract This chapter summarizes the available stratigraphic, petrographical and mineralogical evidence from sediments and sedimentary rocks on the evolution of the Himalayan belt and its associated foreland basin. The use of compositional signatures of modern sediments to unravel provenance changes and palaeodrainage evolution through time is hampered by a poor match with detrital modes of ancient strata markedly affected by selective chemical dissolution of unstable minerals during diagenesis. Only semi-quantitative diagnoses can thus be attempted. Volcanic detritus derived from Transhimalayan arcs since India–Asia collision onset at c. 60 Ma was deposited onto the Indian lower plate throughout the Protohimalayan stage, with the exception of the Tansen region of Nepal that is characterized by quartz-arenites yielding orogen-derived zircon grains. During the Eohimalayan stage, begun in the late Eocene when most sedimentation ceased in the Tethys Himalayan domain, low-rank metasedimentary detritus was overwhelming in the central foreland basin, where a widespread unconformity developed spanning locally as much as 20 myr. Volcanic detritus from Transhimalayan arcs remained significant in northern Pakistan. Arrival of higher-rank metamorphic detritus since the earliest Miocene, and the successive occurrence of garnet, staurolite, kyanite and finally sillimanite, characterized the Neohimalayan stage, when repeated compositional changes in the foreland-basin succession document the stepwise propagation of crustal deformation across the Indian Plate margin and widening of the thrust belt with exhumation of progressively more external tectonic units. The correspondence in time between the activity of major thrusts and petrofacies changes indicates a promising approach to accurately reconstruct the geological evolution of the coupled orogen–basin system. Conversely, a poor conceptual framework and the general reliance on ad hoc mechanisms to explain phenomena unpredicted by simplified models represent major factors limiting the robustness of palaeotectonic interpretations. Improved knowledge requires taking into full account the dynamic role played by still poorly understood subduction processes – rather than exclusively the effect of passive loading – as well as the role played by the presence of inherited structures on the downgoing Indian Plate, which control both lateral variability of orogenic deformation and the location of depocentres in the foreland basin.