Weathering of rocks and its impact on the atmospheric carbon budget have been calculated from chemical compositions of large rivers. Here we present chemical compositions and magnesium isotope ratios for the dissolved and suspended loads of the rivers Danube, Elbe, and Rhine, and investigate whether magnesium isotopes can contribute to the quantification of weathering rates in their catchments. The d26Mg of the dissolved and solid loads vary from -0.93 to -1.85 ‰ and -0.98 to +0.01 ‰ relative to the reference material DSM3, respectively. Although these rivers run through highly populated and industrialized regions, the d26Mg values mirror the lithologies of the catchment areas: the Danubian catchment is dominated by carbonatic lithologies and in the Danube dissolved magnesium exhibits the most negative d26Mg values between -1.85 and -1.70 ‰. The mainly siliceous catchment of the river Elbe causes uniform d26Mg values of -1.00 ± 0.07 ‰ which is within the analytical reproducibility of ± 0.08 ‰ (2sd). During the course of the river Rhine the predominant lithology changes from carbonates to silicates. Consequently, the d26Mg values of dissolved loads rise from -1.74 to -1.36 ‰. Earlier studies found typical d26Mg values of < -2 ‰ for carbonates and values near 0 ‰ for silicates. Therefore, the variation of the magnesium isotopic composition of the studied rivers can be explained by varying amounts of magnesium derived from dissolution of carbonates and alteration of silicates in their catchments. Simple mass balance calculations on the basis of the magnesium isotopes obtained for the investigated rivers imply 26 to 59 % magnesium derived from carbonatic lithologies and 41 to 74 % magnesium from siliceous lithologies. This is in contrast to estimates using conventional methods for the tribute of source lithologies. Our calculation is based on d26Mg values of -0.2 ‰ for silicates and -3.0 ‰ for carbonates. A smaller proportion of siliceous rocks in the dissolved load would be obtained by assuming a higher d26Mg value for the siliceous endmember. However, it is well established now that magnesium of Bulk Silicate Earth is close to -0.2 ‰ and dissolved magnesium from silicates is isotopically light. Therefore, this discrepancy is unlikely solved by assuming an isotopically heavier siliceous endmember. A smaller silicate proportion can also be obtained by changing the assumed d26Mg value of the carbonate endmember, which indeed entails some uncertainty. There is now some evidence that magnesium fractionation during carbonate precipitation is close to -2.5 ‰, giving a d26Mg of -3.3 ‰ for carbonates that precipitated from seawater of -0.8 ‰ relative to DSM3, but dolomite has substantially heavier magnesium and its contribution to the carbonate endmember is unclear. Therefore, direct estimation of the source of dissolved magnesium based on magnesium mass balance may become a powerful tool to quantify weathering fluxes but needs better knowledge of the carbonate endmember composition.