A redshift survey of IRAS galaxies. IV. The galaxy distribution and the inferred density field

We have completed a full-sky flux-limited redshift survey of galaxies detected by the Infrared Astronomical Satellite (IRAS), which we use to characterize the density distribution of galaxies to a redshift of 8000 km s^-1. The galaxy distribution is dominated by the Hydra-Centaurus-Pavo-Indus supercluster (the "Great Attractor") and the Perseus-Pisces supercluster. The Local Supercluster is an extension of the former. There is a large void in front of the Perseus-Pisces supercluster, as well as off the supergalactic plane. Monte Carlo experiments show that we can determine the fractional density field to an accuracy of 0.3-0.4 of the mean density over most of the volume to 8000 km s^-1. We make a detailed comparison of the IRAS density distribution with that inferred from redshift surveys of optically selected samples of galaxies. The IRAS density field agrees well with that defined by the optically selected Center for Astrophysics redshift survey (CfA), except within the densest regions of the richest clusters. For example, the overdensity of galaxies within 200 km s^-1 of the center of the Virgo Cluster is a factor of 2 less in the IRAS than in the CfA survey. A similar discrepancy is seen in the Coma Cluster, although at a lower confidence level. Outside of clusters, the galaxy distributions in the two surveys, when smoothed on scales larger than 500 km s^-1, are consistent with having been drawn from the same density field. There is, however, a localized region beyond the Ursa Major Cluster which shows significantly larger overdensity in the IRAS than in the CfA sample. The clusters follow an r^-2 density profile in both the IRAS and the CfA density fields. Similar conclusions are reached in comparisons with the Southern Sky Redshift Survey and the Supergalactic Plane Survey. The fractional overdensity in IRAS of the sphere centered on the Virgo Cluster with the Local Group on its edge is 1.4. If the IRAS survey is restricted to the CfA solid angle, however, the calculated overdensity is 2.2. The CfA sample itself gives an overdensity of 2.7. Although the large-scale distributions of galaxies in the IRAS and CfA surveys are very similar, the correlation function of the latter is systematically stronger on scales below 1000 km s^-1, consistent with a correlation strength of CfA galaxies twice that of IRAS galaxies. This result is almost unaffected when galaxies in rich clusters of galaxies are given greater weight to compensate for the undersampling of IRAS galaxies in those regions, showing that the effect is not dominated by rich clusters. We conclude that the optical and IRAS samples are consistent with having been drawn from the same density field on scales larger than ∼500 km s^-1, and that on smaller scales the optical surveys show greater clustering, an effect not dominated by the observed IRAS deficiencies in clusters.