Since the first applications of structural concrete, the shear behaviour of one-way slabs without transverse reinforcement has been largely investigated. Nevertheless, currently in the scientific community there is no general agreement on the mechanisms of shear failure, on the parameters governing the shear strength and on the predominant shear-transfer actions. Hence, several mechanical models, based on very different hypotheses, and empirical formulations, calibrated on the available experimental results, have been proposed in the last decades. In addition, these experimental results have been traditionally obtained from tests on simply supported beams subjected to point load, whereas in most one-way slabs without transverse reinforcement in practice (foundations, retaining walls, slabs of cut-and-cover tunnels, silos) the boundary and loading conditions are typically different. This thesis has therefore the objective to provide new experimental data on reinforced concrete members without transverse reinforcement tested with different loading and boundary conditions, to increase the understanding on the mechanisms of shear failure and to develop a mechanical model based on the new experimental evidence. In the first part of the thesis, the experimental results of 23 tests on 20 beams without transverse reinforcement subjected to different loading (concentrated or distributed) and boundary conditions (cantilevers, simply supported or continuous beams) are presented. Refined measurement techniques allowed detailed tracking of the development of the crack pattern up to failure. The results show that the location, inclination and kinematics of the critical shear crack play a major role on the shear strength. Moreover, the amount of shear transferred by the various potential shear-transfer actions has been estimated on the basis of the experimental measurements and by using suitable mechanical models for each shear-transfer action. The analyses show that, for slender members, the shear-transfer actions contributing to the shear capacity are the inclination of the compression chord, the residual tensile strength of concrete, the dowelling action and the aggregate interlock, and the latter is the predominant one. For squat members or members in which the critical shear crack develops below the theoretical compression strut, differently, the arching action is predominant. In the second part of the thesis, a mechanical model, consistent with the main ideas of the critical shear crack theory, is presented. The shear force that is transferred through the critical shear crack by the various shear-transfer actions is calculated by integration of simple constitutive laws and a failure criterion is obtained by summing the different contributions. The shear and deformation capacity can thus be calculated by intersection of the failure criterion with a load-deformation relationship. It is shown that the failure criteria obtained by integration of stresses at the crack surface can be approximated by power-law equations. Combining the power-law failure criteria with the load deformation relationship, a closed-form equation has been obtained. The closed-form equation provides almost identical results to the mechanical model and allows for direct design and assessment of existing structures. The accuracy of the mechanical model and the closed-form equation has been checked against a large database, showing a good agreement to the experimental results.