Rhombic cell analysis as outlined in the first paper of the present series is applied to samples of varying depths and liming luminosities of the IRAS/PSCz Catalogue. Numerical indices are introduced to summarize essential information. Because of the discrete nature of the anlaysis and of the space distribution of galaxies, the indices for a given sample must be regarded as each having an irreducible scatter. Despite the scatter, the mean indices show remarkable variations across the samples. The underlying factor for the variations is shown to be the limiting luminosity rather than the sampling depth. As samples of more and more luminous galaxies are considered over a range of some 2 magnitudes (a factor of some 50 in space density), the morpholgy of the filled and empty regions the galaxies define degrades steadily towards insignificance, and the degrading is faster for the filled than the empty region.

An attempt to answer the question 'Can we observe galaxies that recede faster than light ?' led to a re-examination of the notions of time, distance, velocity and redshift as they occur in newtonian physics, special relativity, general relativity and cosmology. A number of misconceptions were uncovered. It was found that, once freed of special relativity preconceptions, the above question is easily and unequivocally answered

A new way of probing the large-scale structure of the universe is proposed. Space is partitioned into cells the shape of rhombic dodecahedron. The cells are labelled ``filled'' or ``empty'' according as they contain galaxies or not. The cell size is so chosen as to have nearly equal numbers of filled and empty cells for the given galaxy sample. Two observables on each cell are definable: the number of its {\em like} neighbors, $n_1$, and a two-suffixed topological type $\tau$, the suffixes being the numbers of its like and unlike {\em neighbor-groups}. The frequency distributions of $n_1$ and $\tau$ in the observed set of filled (empty) cells are then considered as indicators of the morphology of the set. The method is applied to the CfA catalogue of galaxies as an illustration. Despite its limited size the data offers evidence 1) that the empty cells are more strongly clustered than the filled cells, and 2) that the filled cells, but not the empty cells, have a tendency of occurring in sheets. Further directions of development both in theory and application are indicated.

A more clear-cut answer to the title question is, ``Yes'' if the universal expansion started with a big bang; ``No'' if it started infinitely slowly.