Slab breakoff is the buoyancy‐driven detachment of subducted oceanic lithosphere from the light continental lithosphere that follows it during continental collision. In a recent paperDavies and von Blanckenburg[1994] have assessed the physical conditions leading to breakoff by quantitative thermomechanical modeling and have predicted various consequences in the evolution of mountain belts. Breakoff will lead to heating of the overriding lithospheric mantle by upwelling asthenosphere, melting of its enriched layers, and thus to bimodal magmatism. Breakoff will also lead to thermal weakening of the subducted crustal lithosphere, thereby allowing buoyant rise of released crustal slices from mantle depths. In this paper we present a test of this model in the Tertiary evolution of the European Alps. In the Alps, both basaltic and granitoid magmatism occur between 42 and 25 Ma, following the closure of oceanic basins by subduction and continental collision. The granitoids are now well established to result from mixing of basalt with assimilated continental crust. To identify the tectonically crucial origin of the partial mantle melts, we have compiled all published geochemical and isotopic data of numerous mafic dykes occurring throughout the whole Alpine arc. Their trace element and isotopic composition suggests that they have been formed by low‐degree melting of the mechanically stable lithospheric mantle. We see no evidence for melting of asthenospheric mantle. It was thus not decompressed to depths shallower than 50 km. Once initiated, rapid lateral migration of slab breakoff will result in a linear trace of magmatism in locally thermal weakened crust. This explains why all Alpine magmatic rocks intruded almost synchronously along a strike‐slip fault, the Periadriatic Lineament. A compilation of ages from Penninic high‐pressure rocks subducted to depths of up to 100 km shows that subduction took place at circa 55–40 Ma, followed by uplift at 40–35 Ma. From the short time interval between their uplift and the onset of magmatism we infer that both processes have been induced by the breakoff. The slab breakoff model fulfills its predictions in the case of the Alps and therefore supports the assumptions made in the theoretical model on a geological basis. We believe that the characteristic association of magmatic activity with the return of high‐pressure rocks to the surface allows the identification of this process in the Earth's m