Comment on "Proposal for the Measurement of Bell-Type Correlations from Continuous Variables"

In a recent paper [T. C. Ralph, W. J. Munro, R. E. S. Polkinghorne, Phys. Rev. Lett. 85, 2035 (2000)], the authors propose a test for Bell's inequalities based on quadrature measurements for a correlated parametric source. We present here a local hidden variable model which reproduces all the expectation values predicted in the described experimental scheme, and point out that the proposal is based on an inconsistent use of Bell's inequalities.