Abstract
It is demonstrated that the conjunction of the premises of Einstein-Podolsky-Rosen paper with a supplementary assumption lead to validity of an inequality that is sometimes grossly violated by the quantum-mechanical probabilities in the case of real experiments. The supplementary assumption is as follows: For every photon in the state λ, the sum of the detection probabilities in the ordinary and in the extraordinary beams emerging from a two-way polarizer is less than or equal to the detection probability with the polarizer removed. This assumption, which always holds for an ensemble of photons, is weaker and more general than the assumption of Garuccio and Rapisarda [Nuovo Cimento A 65, 269 (1981)] and the assumption of Clauser and Horne [Phys. Rev. D 10, 526 (1974)]. Hence an experiment based on this assumption refutes a larger family of hidden-variable series than an experiment based on the assumption of Garuccio and Rapisardo or on the assumption of Clauser and Horne.