Many investigations in the USSR and elsewhere have indicated that the discharge of loose materials from tall vessels {silos) is accompanied by local rises in the loads on the walls, pulsation of the loads, and nonuniform loading round the perimeter of the vessel [1][2][3]. Although much effort has been made to elucidate the mechanism of formation of these features, so far there are no satisfactory explanations of the development of the high peak loads or of the pulsations. The rise of pressure during discharge is associated with hanging (arch formation) [4][5][6], with changes in the coefficient of friction of the material on the wall [1], and even with the influence of acoustic shock waves [7]; the asymmetric pressures on the walls are associated with inhomogeneous stacking of the material and nonuniform rigidity of the structure round its perimeter [8].Experts consider that while calculations of the static pressures are rather indeterminate, calculations of the loads arising during discharge of the materials are even more so [9]. This is why there are relatively frequent cases of serious damage to structures for the storage and treatment of loose materials [8,9]. Many investigators have remarked that the solution to this problem will involve further detailed studies of the processes occurring during movement of materials in vessels [10].in [12,13] we showed that the principal disturbances in material flowing out of a vessel involve its motion in convergent channels, i.e., in the conditions of maximally confined deformation. Such channels are components of any bunkers, whether the latter have a convergent section or a fiat bottom; in the latter case the convergent channel is formed within the material itself. Let us therefore first consider experiments on the deformation of loose materials in convergent axisymmetric channels.The experimental methods and the characteristics of the materials used in our investigations were described in detail in [11,12]. To study the kinematics of motion of material in axisymmetric vessels, there is an additional method, ngelatin freezing, n which has proved to be more convenient and less troublesome than the ~paraffin coat freezing n method. Before forming the load, all the material (both light and black sand) was heated to 100*C. After a given portion of the loose material had been discharged through the outlet, the deformed material remaining in the vessel was filled with a heated gelatin solution {water and gelation in the proportions of 10 : 1 by weight). The specimen was then cooled, removed from the vessel, and cut up along various planes.As the convergent channels we used glass funnels and also funnels to which various grades of emery paper had been glued on the inside. The angles of divergence of the funnels were 45-90 ~ and the diameters of the outlets 5-16 ram. The use of glass funnels not only reduced the coefficient of friction between the material and the wall, but also had the advantage that it was possible to make direct observations on the kinematics near the wall...