Before a dry snow slab avalanche is released, a shear failure along a weak layer or an interface has to take place. This shear failure disconnects the overlaying slab from the weak layer. A better understanding of this fracture mechanical process, which is a key process in slab avalanche release, is essential for more accurate snow slope stability models. The purpose of this work was to design and to test an experimental set-up for a mode II fracture test with layered snow samples and to find a method to evaluate the interfacial fracture toughness or alternatively the energy release rate in mode II. Beam-shaped specimens were cut out of the layered snow cover, so that they consisted of two homogeneous snow layers separated by a well defined interface. In the cold laboratory 27 specimens were tested using a simple cantilever beam test. The test method proved to be applicable in the laboratory, although the handling of layered samples was delicate. An energy release rate for snow in mode II was calculated numerically with a finite element (FE) model and analytically using an approach for a deeply cracked cantilever beam. An analytical bilayer approach was not suitable. The critical energy release rate G_c was found to be 0.04±0.02 J m^-2. It was primarily a material property of the weak layer and did not depend on the elastic properties of the two adjacent snow layers. The mixed mode interfacial fracture toughness for a shear fracture along a weak layer estimated from the critical energy release rate was substantially lower than the mode I fracture toughness found for snow of similar density.