In the Netherlands, the shear capacity of a large number of existing reinforced concrete solid slab bridges is subject to discussion, as initial assessments indicated that their capacity was insufficient. In certain cases, the deterministic value of the moment capacity is larger than the deterministic value of the shear capacity. However, when the variability of the material properties, and of the capacity models themselves are factored in, a probability of a certain failure mode can be calculated. Here, a method is introduced to calculate the chance that a cross-section fails in shear before it fails in bending. The method that is derived here is applied to the Ruytenschildt Bridge. This case study is a reinforced concrete solid slab bridges that was tested to failure in two spans during the summer of 2014. The relative probability of failure in shear of the bridge was determined. The predictions indicated a smaller probability of a shear failure than of a bending moment failure. In the first tested span, failure was not reached, but indications of flexural distress were observed. In the second span, a flexural failure was achieved, in line with the probabilistic predictions. The presented method can be used in the assessment of existing bridges to determine which failure mode is most probable, taking into account the variability of materials and capacity models. Comments: The reviewer believes that there are still two vital problems in this manuscript: 1) Using only two tests (and one of them did not achieve failure) to prove the proposed method is not sufficient. And if a method is not verified, it could not be used in any way. The authors responded: "The authors do not think comparing the method to laboratory tests would give interesting results, because typical shear tests are overdesigned in bending to make sure that the element fails in shear. Similarly, experiments on beams in bending will be designed so that the shear span is large enough to avoid a shear failure, or beams for flexural tests can be reinforced with stirrups to avoid a shear failure". However, I think that laboratory tests would still give some benefits to the verification of the proposed method. For example, although typical beam or slab shear tests are overdesigned in bending to make sure that the members fail in shear, the ratios of the bending capacity to shear capacity of the test members are different for each test, which means that the predicted probability of shear failure would be different. Furthermore, it could be inferred that if the ratio of the bending capacity to shear capacity is getting smaller, the probability of shear failure is also getting smaller. If the method proposed by the authors could predict this tendency, it could be stated that the method was verified reasonably. Actually, it is the core value of the probabilistic method to obtain a specific failure probability regardless of how a member is designed. I've verified the method with four representative laboratory tests: S1T1, S5T4, S8T1 and S9T1: Case Failure mo...