The aim of this study was to give some new constrains on the Himalayan erosion, using the geochemistry of dissolved and particulate matter carried by Himalayan rivers. The main results are: - The supended load is not representative of the total erosion load. For the Ganges-Brahmaputra (G-B), it represents less than 50 % of the total particulate discharge. The total erosion in the G-B is therefore 2.5 ± 0.8 10.9 ton/yr. - The location of intense erosion corresponds to the location of heavy rain in the range or at the G-B scale. The erosion rate are 3.4 ± 0.7 mm/yr. for the Brahmapulra Himalaya and 2.3 ± 0.6 for the Ganges Himalaya while the runoff of these 2 domains are 2.1 and 1.2 m/yr., respectively. In central Nepal, erosion rates follow the climatic contrast between the south and the North flank of the range: the erosion of the South flank of the high range is more than 6 limes higher than that of the North flank. - The highest erosion rate is located just above the crustal ramp of the main Himalayan thrust, bellow the heaviest rain and on glaciated terranes. Tectonic convergence and the geometry of the crustal thrust focus the vertical uplift in the narrow zone of the south flank of the high range, where the erosion is twice more important than everywhere else. - The Himalayan erosion is buffered by physical processes. For the G-B, more than 55% of the eroded carbonate are carried by the particulate phase and no more than 1.3% of the eroded silicate have been dissolved. - The high erosion rate acts as a restriction of the weathering. ln the south flank of the high range, where the erosion rate is the highest, the weathering of rocks is the lowest. This result implies that there is not a general and positive relationship between tectonic and weathering. This link is even more camplex for the whole orogenic process since the weathering of highly divided material in the foreland basin has to be taken into account. - The alkalinity flux corresponding to silicate weathering is 2.7 10.11 mol/yr. corresponding to 2.3% of the global flux. The budget of the weathering implies a dominant contribution of the carbonate dissolution to the riverine chemistry and that weathered silicate are mostly sodic. The long term atmospheric CO2 consumption by silicate weathering has been estimated to be 6,4 10.10 mol/yr. for the whole G-B. - 14 Eq% of the alkalinity corresponds to an abiotic alteration. The main is the sulfide oxidation that produces around 70% of the dissolved sulfate load of the G-B. Some metamorphic CO2 degassing, along Tibetan graben account for less than 1 % of the whole alkalinity, but is locally significant. - Dissolved Sr budget and si licate alkalinity budget are uncoupled. An incongruent dissolution of carbonate in the Tibetan part of the range is the main source of dissolved Sr while the small weathering of old crustal terranes (> 2Ga) on the south flank acts as a 87Sr spike for the Himalayan rivers. The G-B (6.5 10.8 mol/yr. of dissolved Sr with 87Sr/86Sr -0.730) plays a key role of the Sr oceanic budget but has a few impact on the alkalinity. - The global impact on the carbon cycle by the Himalayan erosion lies in its ability to preserve and bury organic carbon. This burial is one order of magnitude higher than the long term atmospheric CO2 consumption by silicate weathering and represents 5% of the global C burial. This enhanced C removal from the ocean-atmosphere reservoir by the erosion of the Himalaya couId be suffïcient to deslabilize the global carbon cycle.