Voids are the most prominent feature of the LSS of the universe. Still, they have been generally ignored in quantitative analysis of it, essentially due to the lack of an objective tool to identify and quantify the voids. To overcome this, we present the Void-Finder algorithm, a novel tool for objectively quantifying galaxy voids. The algorithm classifies galaxies as either wall- or field-galaxies. Then it identifies voids in the wall-galaxy distribution. Voids are defined as continuous volumes that do not contain any wall-galaxies. The voids must be thicker than an adjustable limit, which is refined in successive iterations. We test the algorithm using Voronoi tessellations. By appropriate scaling of the parameters we apply it to the SSRS2 survey and to the IRAS 1.2 Jy. Both surveys show similar properties: ~50% of the volume is filled by the voids, which have a scale of at least 40 Mpc, and a -0.9 under-density. Faint galaxies populate the voids more than bright ones. These results suggest that both optically and IRAS selected galaxies delineate the same LSS. Comparison with the recovered mass distribution further suggests that the observed voids in the galaxy distribution correspond well to under-dense regions in the mass distribution. This confirms the gravitational origin of the voids.

Submitted to ApJ; 33 pages, aaspp4 LaTeX file, using epsfig and natbib, 1 table, 12 PS figures. Complete gzipped version is available at http://shemesh.fiz.huji.ac.il/hagai/; uuencoded file is available at http://shemesh.fiz.huji.ac.il/papers/ep3.uu or ftp://shemesh.fiz.huji.ac.il