Abstract. We discuss the experimental techniques used to date for measuring the changes
in polarization state of a laser produced by a strong transverse magnetic field
acting in a vacuum. We point out the likely artifacts that can arise in such
experiments, with particular reference to the recent PVLAS observations and the
previous findings of the BFRT collaboration. Our observations are based on
studies with a photon-noise limited coherent homodyne interferometer with a
polarization sensitivity of 2x10^-8 rad Hz^(1/2) mW^(-1/2).