Many radars around the world are currently adapting the simultaneous horizontal (H) and vertical (V) polarization transmit technique to achieve dual polarization measurements. The technique is based on the observation that the mean canting angle of large ensembles of raindrops is zero degrees. This implies that there is no cross-coupling of the H and V transmitted signals. It is know that if the mean canting angle of the precipitation particles is non zero, then there will be cross-coupling and biases will become evident in the polarimetric signatures. The ice phase and the melting/freezing layer of storms are regions where this can occur. Radar observations have also supported and shown this, e.g., Ryzhkov and Zrnic, “Depolarization in Ice Crystals and Its Effect on Radar Polarimetric Measurements”, JTECH, 2007. Similar bias effects can also be caused by polarization errors of the radar system. Polarization errors will occur because of non-ideal antenna and feed horn, support struts and other RF devices. This paper examines the possible error sources that effect the polarization state of the transmitted wave. A model is developed and the magnitudes of the errors are quantified with particular attention to the measurement of the differential reflectivity, Zdr. The model demonstrates how the errors are a strong function of the differential propagation phase, phi_dp. The effect of the differential phase between the H and V transmit components is also examined. Estimation of radar polarization errors is discussed and data from NCAR's S-band polarimetric radar, S-Pol, is used to illustrate some of the concepts.