Hubble (1929) presented evidence for a linear relationship between the distance to a galaxy estimated from absolute magnitude criteria and the observed velocity redshift cz of the galaxy. In the standard interpretation of the Hubble ?law?, deviations from a linear Hubble relationship are relatively small ? no more than 1500 km s-1 ? and are thought to result from bulk flows and peculiar motions gravitationally induced by large mass concentrations. Evidence has accumulated which indicates that some galaxies may deviate significantly from the linear Hubble flow ? possibly by over 10,000 km s-1 in the most extreme cases. In many instances, evidence for interaction ? including bridges - apparently connect two galaxies and sometimes quasars with very different redshifts. In these cases of ?discordant redshift?, the bridges are generally not considered to be real connections but instead are viewed as accidental projections of a foreground object with a higher redshift background object. However, if the bridges are real connections between objects with large redshift differences, then they would support the existence of a non-cosmological redshift component in some galaxies and quasars.
The best way to establish the existence of a non-cosmological redshift is to determine distances to galaxies using redshift independent distance indicators with small enough scatter to rule out large errors in the calculated distance. In this analysis the Ks-band Tully-Fisher relationship is used to determine distances to normal spiral galaxies. It is found that a number of spiral galaxies have deviations from Hubble's law ranging from 2500 km s-1 to 4000 km s-1. Several tests and comparisons demonstrate that these deviations are unlikely to be false deviations resulting from large Tully-Fisher distance errors. Therefore these excessively large deviations from the Hubble law support the possible existence of a non-cosmological redshift component in some normal spiral galaxies.