Interpretation of the Global Anisotropy in the Radio Polarizations of Cosmologically Distant Sources

We present a detailed statistical study of the observed anisotropy in radio polarizations from distant extragalactic objects. This anisotropy was earlier found by Birch (1982) and reconfirmed by Jain and Ralston (1999) in a larger data set. A very strong signal was seen after imposing the cut $\mathrm{\mid }RM-\overline{RM}\mathrm{\mid }>6$ rad/m${}^2$, where $RM$ is the rotation measure and $\overline{RM}$ its mean value. In this paper we show that there are several indications that this anisotropy cannot be attributed to bias in the data. We also find that a generalized statistic shows a very strong signal in the entire data without imposing the RM dependent cut. Finally we argue that an anisotropic background pseudoscalar field can explain the observations.