Historically, the first boundary conditions to be formulated and used in the theory of ferromagnetic thin films, the Rado-Weertman (RW) conditions, have a general advantage of being a simple differential equation, 2Aex ∂m ∂n − K surf m = 0. A key role in this equation is played by the phenomenological quantity K surf known as the surface anisotropy energy density; Aex denotes the exchange stiffness constant, and m is the amplitude of the transverse component of dynamic magnetization. In the present paper we use a microscopic theory to demonstrate that the surface anisotropy energy density of a thin film is directly related with its free-energy density, a fact not observed in the literature to date. Using two local free-energy densities F surf and F bulk , defined separately on the surface and in the bulk, respectively, we prove that K surf = d F surf − F bulk , where d is the lattice constant. The above equation allows to determine the explicit configuration dependence of the surface anisotropy constant K surf on the direction cosines of the magnetization vector for any system with a known formula for the free energy. On the basis of this general formula the physical boundary conditions to be fulfilled for a fundamental uniform mode and surface modes to occur in a thin film are formulated as simple relations between the surface and bulk free-energy densities that apply under conditions of occurrence of specific modes. The Rado-Weertman (RW) boundary equation is the earliest boundary condition to have been proposed in the theory of thin-film magnetism [1,2] with the aim of taking into account the specific dynamics of motion of the surface magnetization, as distinct from that of the bulk magnetization. The RW equation is based on a continuum model of magnetization and in the circular approximation readswhere m is the amplitude of the transversal (dynamic) component of the magnetization, A ex is the exchange stiffness constant, n denotes the direction normal to the surface of the film, K surf (ϑ, ϕ) is the surface anisotropy energy density, and magnetization angles ϕ and ϑ are azimuth and polar angles, respectively. The RW Eq. (1) is mostly used for the description of two extreme situations, defined by the conditions:The first condition (2) implies zero surface anisotropy constant, K surf (ϑ, ϕ) ≡ 0; this is considered to correspond to completely unpinned surface spins. The fulfillment of the other condition (2) requires that |K surf (ϑ, ϕ) | → ∞, and an infinite value of the surface anisotropy constant is considered to result in a situation in which surface spins are completely pinned. * e-mail: henpusz@amu.edu.plWe will now show that the differential form of the RW equation written above is equivalent to the microscopic boundary equation derived within the surface inhomogeneity (SI) model; the latter is based on a discrete model of magnetization and has the form of difference equation [3,4]:where m 0 ≡ m surf is the surface amplitude of the transversal (dynamic) magnetization component m, and m −d is its analytic...