Cosmogenic isotopes, such as 3He, l~ 26A1, are continuously producted at the Earth's surface in rocks and soils by cosmic ray spallation of light isotopes, providing a powerful tool for dating young geomorphological processes. The datation method based on the accumulation of cosmogenic 3He is particulary appealing because: 3Hec (where suffix c refers to 3He produced by cosmic rays) is a stable isotope, its production rate is high and its natural background is very low. Moreover, its analysis is sensitive, relatively simple and rapid. However, the applicability of this method is limited by the high diffusivity of 3He. In practice, 3He~ is quantitatively retained in olivine phenocrysts, allowing datation of young basaltic flows. Dung/ and Roselieb (1996) have argued that the diffusivity of helium is lower in garnet than in olivine. The use of garnet for 3He c dating presents the interesting advantage to extend the field of this method to crustal metamorphic domains, and therefore to the case of major orogens. We have measured3Hec in alluvial garnets sampled in bedload from Central Nepal rivers (Fig. 1). They represent 3 different monolithologic catchments. By analysing fractions of several tenths of grains, we integer the complex history of each individual grain and the results obtained are mostly meaningful for comparison between different positions in a river basin. The garnets were formed in gneisses of the High Himalaya Crystalline (HHC) and schists of the Lesser Himalaya (LH) during high grade metamorphism resulting from the Main Central Thrusting (MCT) during Miocene. In this system, uplift rates estimated using the Ar/Ar method on micas, feldspars and monazite in garnet are high between 2 and 4 mm/yr depending on the position with regard to the MCT (Copeland et al., 1991). For 3Heo dating purpose, pyropes and almandines are more appropriate than grossularites et spessartites (Dung/and Roselieb, 1996). Electron probe analysis showed that the mean composition of the garnets was Aim. 70%, Pyr. 15%, Gros. 7%, Spes. 5%, Andr. 3% except for sample MO38 wich showed a larger contribution of the grossular end-member (Alm. 54%, Pyr. 12%, Gros. 26%, Spes. 6%, Andr. 2%). Helium was extracted using a CO2 laser from mgsized pure garnet fractions and its amount and isotopic ratio measured with a high sensitivity rare gas mass spectrometer in Nancy (SHe = 3.5 • 10 -4 A/Torr, He blank = 2 x 10 15 mol). Helimn-3 present in minerals is the sum of the following potential contributions: 3Her 3He c + 3Hea + 3Hen + 3He m (where suffixes t, a, n and m refer to total, atmospheric, nucleogenic, and mantle-derived). The parent rocks of metamorphic garnets are thought to be marine sediments, mostly turbidites, which makes the possible contribution of mantle-derived 3He m unlikely. In agreement with that, He isotopic ratios measured in hot springs along the MCT are show typical radiogenic values (Marty et al., 1996). 3Hea was estimated from the measured 2~ content. The contribution of nucleogenic 3He in garnet, produced by activation of Li by natural thermal neutrons, was computed from ion probe analysis of Li in garnet and from U and Th analysis &garnets (1CP-MS), or from estimate of U, Th content of parent rocks. In all cases the atmospheric and nucleogenic 3He contributions were found to be small with respect to the total 3He content of the samples.