Abstract
The correlation energy computed from Peebles' power-law galaxy covariance function suggests that, if the gravitational instability picture is correct, then there must be a feature in the covariance function where it steepens. The scale of this cut-off cannot be accurately determined by this argument but a comparison of the random kinetic energies of galaxies with their correlation energy gives an estimate of the mean matter density of the Universe. The suggested value is less than 5 per cent of the closure density if most matter is distributed as galaxies are.