Experimental observations of capillary rise and hysteresis of water or ethanol in glass beads are presented to improve our understanding of those physical processes in porous media. The results provide evidence that capillary rise into porous media cannot be fully explained by a model of cylinders. They further demonstrate that the ''ink bottle'' model does not provide an adequate explanation of hysteresis. Glass beads serving as a model for ideal soil are enclosed in a rectangular glass chamber model. A TV camera associated with a microscope was used to record the processes of capillary rise and drainage. It is clearly shown during capillary rise that the fluid exhibits a ''jump'' behavior at the neck of the pores in an initially dry profile or at the bottom of the water film in an initially wet profile. Under an initially dry condition, the jump initiates at the particle with smallest diameter. The jump process continues to higher elevations until at equilibrium the surface tensile force is balanced by the hydrostatic force. The wetting front at that time is readily observed as flat and saturated. Under an initially wet condition, capillary rise occurs as a water film thickening process associated with the jump process. Trapped air behind the wetting front renders the wetting front irregular and unsaturated. The capillary rise into an initially wet porous medium can be higher than that into an initially dry profile. During the drying process, large surface areas associated with the gas-liquid interface develop, allowing the porous medium to retain more water than during the wetting process at the same pressure. That mechanism explains better the hysteresis phenomenon in porous media in contrast to other mechanisms that now prevail. ¿ American Geophysical Union 1994