A computation of the covariance between two linear statistics for the Jellium model

Abstract

We extend previous results providing an exact formula for the variance of a linear statistic for the Jellium model, a one-dimensional model of Statistical mechanics obtained from the $k \longrightarrow 0^{+}$ limit of the Dyson log-gas. For such a computation of the covariance, in comparison to previous work for computations of the log-gas covariance, we obtain a formula between two linear statistics, given arbitrary functions $f$ and $g$ over the real line, that is dependent upon an asymptotic approximation of the Jellium probability distribution function from large $N$ deviations, and from an effective saddle-point action.