A CCD study of the environment of Seyfert galaxies. II. Testing the interaction hypothesis

An analysis of the environment of a sample of 33 CfA Seyfert galaxies and a control sample of 45 nonactive galaxies matched in luminosity, redshift, and morphology to the Seyfert galaxies as reported in Paper I is presented. The covariance function amplitudes of the Seyfert and control samples are not statistically significantly different from one another and from the general held. Moreover, the companion frequency of the Seyfert galaxies, the probability of finding a companion galaxy brighter than -17.5 in R within 50 kpc (0.30 +/- 0.11), is not statistically significantly different from that-for the nonactive control sample (0.23 +/- 0.09). The mean environment of Seyfert 1 galaxies is found to be different from that of Seyfert 2 galaxies at greater than the 95% confidence level, in the sense that the latter have a larger covariance amplitude. Such evidence is problematic for the Unified Model which attributes spectroscopic differences between the classes to purely geometric effects on the order of parsec scales. It cannot, however, account for differences on the order of 100 kpc scales. It is argued that triggering of activity in galactic nuclei may involve a variety of mechanisms and may depend on the luminosity of the class. That is, while there is excellent evidence that QSOs, radio galaxies, and BL Lac objects inhabit environments significantly richer than the held the same does not seem to be true for Seyfert galaxies and perhaps for LINERs. Finally, because a significant fraction of Seyfert host galaxies show little or no evidence for a recent merger, it is suggested that "minor mergers," mergers that involve a gas-rich disk galaxy and a bound companion or satellite galaxy, may play a significant role in triggering activity in Seyfert galaxies.