The observational status of the effect of bars on AGNs is a bit more confusing. It is sometimes said in the literature (e.g., Shlosman et al. 1989) that all AGNs are barred; however, the evidence is far from clear. Adams (1977) and Simkin et al. (1980) first remarked that a large fraction of Seyfert galaxies contain morphological features such as rings and bars, suggesting that the latter might be instrumental in delivering fuel to the active nuclei. It must be remembered, however, that both of these early studies were based on rather small samples of galaxies, and that only a slight preponderance of barred host galaxies was suggested. Using a larger sample of objects, Arsenault (1989) reported an overabundance of barred galaxies with inner rings among H II nuclei and AGNs. However, as discussed in Ho, Filippenko, & Sargent (1996b), Arsenault's analysis is potentially flawed due to his particular choice of the control sample. Márquez (these proceedings) and Moles, Márquez, & Pérez (1995) did not find an excess of barred galaxies in a sample of AGNs selected from the Véron-Cetty & Véron (1991) catalog, although they suggested that there was an excess of inner rings among the subsample of AGNs lacking bars. It is difficult to assess the significance of the latter result without a proper control sample. As noted by Moles et al., contrary to the classification of bars, that of inner rings is not always consistently applied in the RC3 catalog (de Vaucouleurs et al. 1991). Moreover, unlike the case of most bars, the small linear size of inner rings makes their identification more dependent on angular resolution, and hence subject to more complicated selection effects.
A number of studies, albeit based on rather small samples, suggest that bars have an insignificant effect on AGNs (Heckman 1980; Fricke & Kollatschny 1989; MacKenty 1990). McLeod & Rieke (1995) examined the morphological properties of two samples of Seyfert galaxies - one distilled from the CfA redshift survey (Huchra & Burg 1992) and the other drawn from bright, nearby galaxies (Maiolino & Rieke, unpublished). They did not find an excess of barred galaxies in both samples. Tsvetanov (these proceedings) reported similar results from a sample of southern Seyferts. It has been argued (e.g., Shlosman, Begelman, & Frank 1990; Heckman 1992) that the fraction of barred galaxies may be underestimated in optical catalogs , as the effect of dust obscuration is strong and the old stellar population does not have a large contrast in visible light. Although near-IR imaging surveys sometimes do find previously unrecognized bars (e.g., Keel, Byrd, & Klaric, these proceedings), it seems quite clear that bars are not universally present in AGNs (McLeod & Rieke 1995).