In typical reinforced concrete design, reinforcement is designed to carry axial forces, but it can also resist transversal forces by dowel action. This is usually neglected for simplicity's sake in the design phase, but it can be accounted for either explicitly in mechanical models or implicitly in empirical relationships. Furthermore, there are cases where the connection between various concrete elements explicitly depends on dowel action, as for example, in connections between precast elements or between two concrete parts cast at different times. On the other side, dowel action can have a negative impact on the fatigue resistance of reinforcing bars subjected to cyclic loading, because of the local stress concentrations near interfaces due to relative movements, either in sliding or in opening of cracks not perpendicular to the bar. For the assessment of the remaining capacity of existing structures, improved models of the behavior are needed, including realistic models of the behavior of concrete, steel and their interfaces. The aim of the present paper is to provide a contribution to a better understanding of dowel action by two test series. The first series focused on the behavior of the dowel: the concrete specimens with the embedded bars were placed in a custom-made test setup and subjected to monotonic or low stress-level cyclic actions with a longitudinal and a transversal crack opening component, up to developing the full plastic capacity of the dowel and rupture at the peak of catenary action. The measurement system included tracking the displacement field at the surface of the concrete and the strains in the dowel by optical fibers glued on its surface. The latter measurements allow to derive the internal forces in the reinforcing bar and deformed shape of the bar as well as the contact pressure between the bar and the surrounding concrete. The results show a strong dependency on the test variables: diameter of the bar, imposed crack kinematics and angle between the bar and the crack. The second test series looked more closely at the behavior of concrete underneath the bar, in the presence of a point load introduced at various locations into concrete through a reinforcing bar. A comparison of the test results with existing models shows a general good agreement and some aspects that deserve to be improved.