[en] Geller and Peebles (Astrophys J.; 184:329(1973)) have devised a statistical method for analysing nearby groups of galaxies which is meant to replace the time average of the virial theorem with an ensemble average over groups. The basis of their method is examined and is shown to be essentially equivalent to the correlational approach outlined by Fall who used the method to set limits on the mean mass density of clustered matter in the Universe. If galaxies have clustered by the mechanism of gravitational instability and if most galaxies deviate from Hubble flow by less than about 1200 km s^-1 then the suggested upper limit for clustered matter is less than the Universe's closure density. With Geller and Peebles velocity data, the indicated density of clustered matter is less than 5 per cent of the closure density. (author)