The interaction of microwaves with the natural snow cover strongly depends on the complex structure of the snowpack. In order to quantify this dependency, dedicated experiments were performed with homogeneous slabs of dry, natural snow samples measured over a frequency range from 11 to 94 GHz. A new method introduced by Mätzler and Wegmüller [1995] and Weise [1996a] for determining the scattering and absorption behavior of test samples was applied and further developed by application of a multiple scattering model. Homogeneous samples of dry snow were (1) investigated using a set of portable, linearly polarized Dicke radiometers at frequencies of 11, 21, 35, 48 and 94 GHz, (2) characterized by temperature, grain size and shape, density and permittivity, and (3) structurally analyzed by digitized snow sections in order to obtain statistical information of the snow structure i.e. the autocorrelation function. During the winters 1994/1995 and 1995/1996 additional measurements of snow samples were made to extend the variability of the investigated snow types. Up to now, 20 samples, representing alpine snow in winter (that is, without melt metamorphism) have been collected during three winter campaigns. Here, we present the method and the radiative transfer model and show how it can be inverted to obtain scattering and absorption coefficients. A first assessment of the snow sample data is also presented. The results show good agreement between the measured and the theoretical absorption coefficient. The scattering coefficient turns out to be a strong function of frequency and correlation length as expected from Rayleigh scattering. However, distinct differences can be noted.