An analysis of voids using cosmological N-body simulations of cold dark matter models is presented. It employs a robust statistics of voids, that was recently applied to discriminate between data from the Las Campanas Redshift Survey and different cosmological models. Here we extend the analysis to 3D and show that typical void sizes D in the simulated galaxy samples obey a linear scaling relation with the mean galaxy separation lambda: D=D_0+nu*lambda. It has the same slope nu as in 2D, but with lower absolute void sizes. The scaling relation is able to discriminate between different cosmologies. For the best standard LCDM model, the slope of the scaling relation for voids in the dark matter halos is too steep as compared to the LCRS, with too small void sizes for well sampled data sets. The scaling relation of voids for dark matter halos with increasing mass thresholds is even steeper than that for samples of galaxy-mass halos where we sparse sample the data. This shows the stronger clustering of more massive halos. Further, we find a correlation of the void size to its central and environmental average density. While there is little sign of an evolution in samples of small DM halos with v_{circ} ~ 90 km/s, voids in halos with circular velocity over 200 km/s are larger at redshift z = 3 due to the smaller halo number density. The flow of dark matter from the underdense to overdense regions in an early established network of large scale structure is also imprinted in the evolution of the density profiles with a relative density decrease in void centers by 0.18 per redshift unit between z=3 and z=0.

We have analyzed the size distribution of voids in 2D slices of the Las Campanas Redshift Survey (LCRS). The characteristic sizes show a scaling $D = D_0 + \nu\lambda$ with the mean galaxy separation $\lambda$. Comparison with mock samples of the LCRS in 2D and 3D, using various simulations, cosmologies and galaxy identification schemes, gives a similar scaling, but with steeper slope and a lack of large voids. Best results are obtained for dark matter halos in a $\Lambda$CDM model.

We have analyzed the distribution of void sizes in the two-dimensional slices of the Las Campanas Redshift Survey (LCRS). Fourteen volume-limited subsamples were extracted from the six slices to cover a large part of the survey and to test the robustness of the results against cosmic variance. Thirteen samples were randomly culled to produce homogeneously selected samples. We then studied the relationship between the cumulative area covered by voids and the void size as a property of the void hierarchy. We find that the distribution of void sizes scales with the mean galaxy separation, $\lambda$. In particular, we find that the size of voids covering half of the area is given by $D_{med} \approx \lambda + (12\pm3) \h^{-2}$Mpc. Next, by employing an environmental density threshold criterion to identify mock galaxies, we were able to extend this analysis to mock samples from dynamical $n$-body simulations of Cold Dark Matter (CDM) models. To reproduce the observed void statistics, overdensity thresholds of $\delta_{th} \approx 0 ... 1$ are necessary. We have compared standard (SCDM), open (OCDM), vacuum energy dominated ($\Lambda$CDM), and broken scale invariant CDM models (BCDM): we find that both the void coverage distribution and the two-point correlation function provide important and complementary information on the large-scale matter distribution. The dependence of the void statistics on the threshold criterion for the mock galaxy indentification shows that the galaxy biasing is more crucial for the void size distribution than are differences between the cosmological models.