Maximum Likelihood Comparison of Tully‐Fisher and Redshift Data. II. Results from an Expanded Sample

Abstract

This is the second in a series of papers in which we compare Tully-Fisher (TF) data from the Mark III Catalog with predicted peculiar velocities based on the IRAS galaxy redshift survey and gravitational instability theory, using a rigorous maximum likelihood method called VELMOD. In Paper I (Willick et al. 1997b), we we applied the method to a $cz_{LG} \leq 3000$ km/sec, 838-galaxy TF sample and found $\beta_I=0.49\pm 0.07,$ where $\beta_I\equiv \Omega^{0.6}/b_I$ and $b_I$ is the linear biasing parameter for IRAS galaxies. In this paper we increase the redshift limit to $cz_{LG}=7500$ km/sec, thereby enlarging the sample to 1876 galaxies. The expanded sample now includes the W91PP and CF subsamples of the Mark III catalog, in addition to the A82 and MAT subsamples already considered in Paper I. We implement VELMOD using both the forward and inverse forms of the TF relation, and allow for a more general form of the quadrupole velocity residual detected in Paper I. We find $\beta_I=0.50\pm 0.04$ (1-sigma error) at 300 km/sec smoothing of the IRAS-predicted velocity field. The fit residuals are spatially incoherent for $\beta_I=0.5,$ indicating that the IRAS plus quadrupole velocity field is a good fit to the TF data. If we eliminate the quadrupole we obtain a worse fit, but a similar value for $\beta_I$ of $0.54\pm 0.04.$ Changing the IRAS smoothing scale to 500 km/sec has almost no effect on the best $\beta_I.$ We find evidence for a density-dependence of the small-scale velocity dispersion, $\sigma_v(\delta_g)\simeq (100 + 35 \delta_g)$ km/sec.