The Two--Point Correlation Function and the Size of Voids

Under the assumption of a void-filled Universe we investigate if the characteristic scale of voids can be determined from existing surveys. We use the Voronoi tessellation to create mock surveys and study the properties of the first zero-crossing of the two-point correlation function for various survey geometries. Our main conclusion is that the available data sets should be able to discriminate between 5000 \kms and 12000 \kms voids, if one of these scales actually characterizes the distribution of galaxies.