The scale and dispersion of galactic alignments

The alignment of galaxies is considered using circular statistics to model random tidal torques. It is shown that an initial alignment is not greatly dispersed by subsequent torques and should remain detectable even with quite small samples. It is shown that the dispersion can be used to estimate the amount of random torque experienced since the initial alignment was imposed and hence to estimate the age of the alignment and so of the galaxies themselves. It is suggested that the ambiguity of the existing data may result either from the investigation of clusters and superclusters larger than the typical alignment scale (in the mistaken belief that such large samples are needed to achieve statistical significance), or from the galaxies being older than is generally supposed. A method is described for determining the optimum sample size for use in future studies so as to distinguish between these two possibilities.