5. DEMOGRAPHICS OF EMISSION-LINE GALAXIES

Although the specific numbers cited differ from one investigator to another, all the older surveys discussed in Section 3 agree that LINERs are extremely common in nearby galaxies. They also concur that the detection rate of LINERs varies strongly with Hubble type, with early-type systems being the preferred hosts; this result essentially confirms what was already found by Burbidge & Burbidge (1962), who noted that most of the galaxies showing enhanced [N II] / H ratios tended to be of early type.

Not surprisingly, the Palomar survey likewise finds the same trends. The important distinction, however, is that the results from the present survey are quantitatively much more reliable, for reasons already discussed in Section 4, both in a statistical sense as well as on an object-by-object basis. The detection rates of the various classes of emission-line nuclei defined in Section 2 are given in Table 2 and graphically illustrated in Figure 4a. The conclusions that can be drawn are the following.

Figure 4. (a) Percentage of galaxies with the various classes of emission-line nuclei detected as a function of Hubble type. (b) Distribution of the classes of emission-line nuclei as a function of the absolute B magnitude of the host galaxy.

1. At the limit of our survey, which is at least 4 times more sensitive to the detection of emission lines than any of the older surveys, most galaxies (86%) exhibit optical line emission in their central few hundred parsecs, implying that ionized gas is almost invariably present. This fraction, of course, represents a lower limit. Keel (1983a) detected emission in all the galaxies he surveyed, but his sample was restricted to spirals; Table 2 confirms that essentially all spirals have nuclear emission lines. The Hubble type distribution of the surveys of Heckman et al. (1980) and Véron-Cetty & Véron (1986) more closely matches that of the present sample, and, in these, the detection rate was only ~ 60%-65%.

2. Seyfert nuclei can be found in at least 10% of all galaxies with B_T 12.5 mag, the vast majority of which (~ 80%) have early Hubble types (E-Sbc). The fraction of galaxies hosting Seyfert nuclei has roughly doubled compared to previous estimates (Stauffer 1982b; Keel 1983b; Phillips, Charles, & Baldwin 1983; Maiolino & Rieke 1995). It is interesting to note that Seyfert nuclei, at least with luminosities as low as those here, do not exclusively reside in spirals, as is usually believed (e.g., Adams 1977; Weedman 1977). In fact, galaxies of types E and E/S0 have roughly the same probability of hosting a Seyfert nucleus as those of types between S0 and Sbc.

3. ``Pure'' LINERs are present in ~ 20% of all galaxies, whereas transition objects, which by assumption also contain a LINER component, account for another ~ 15%. Thus, if all LINERs can be regarded as genuine AGNs, they truly are the most populous constituents - they make up > 70% of the AGN population (here taken to mean all objects classified as Seyferts, LINERs, and transition objects) and a full 1/3 of all galaxies. The latter statistic broadly supports earlier findings by Heckman (1980b) and others.
4. The Hubble type distribution of ``pure'' LINERs is virtually identical to that of Seyferts; the same can be said for the distribution of absolute magnitudes (Fig. 4b), both groups having a median M<sub>B</sub> = -20.2 mag. On the other hand, the hosts of many transition objects apparently have somewhat later Hubble types and fainter absolute magnitudes (median M<sub>B</sub> = -20.0 mag), consistent with the idea that these systems are composites of ``pure'' LINERs and H II nuclei.

5. H II nuclei, in striking contrast to AGNs, occur preferentially in late-type galaxies (Heckman 1980a; Keel 1983a; Terlevich, Melnick, & Moles 1987). Quite surprisingly, not a single elliptical galaxy falls into this category. This is consistent with the survey of early-type (E and S0) galaxies of Phillips et al. (1986); the few objects they identified as having H II nuclei are all classified S0 (two are E-S0). Narrow-band imaging surveys of elliptical galaxies (e.g., Shields 1991) often reveal detectable amounts of warm (T 10⁴ K) ionized gas in their centers. Although the dominant ionizing agent responsible for the line emission is still controversial (Binette et al. 1994, and references therein), our failure to detect spectra resembling ordinary metal-rich H II regions among the ~ 60 ellipticals in our survey suggests that young massive stars are probably not the culprit, unless the physical conditions in the centers of ellipticals conspire to make H II regions look very different from those seen in the nuclei of S0s and early-type spirals.

Theoretical studies (e.g., Heller & Shlosman 1994) suggest that large-scale stellar bars can be highly effective in delivering gas to the central few hundred parsecs of a spiral galaxy, which may then lead to rapid star formation. Further instabilities may result in additional inflow to smaller physical scales relevant for AGNs. Thus, provided that a reservoir of gas exists, the presence of a bar might be expected to influence the fueling rate, and hence the activity level. Being sufficiently large and unbiased with respect to bar type, the Palomar survey can be used to examine this issue. Ho, Filippenko, & Sargent (1996a, e) find that the presence of a bar does indeed enhance both the probability and rate of the formation of massive stars in galaxy nuclei, but only for spirals with types earlier than Sbc. By contrast, AGNs seem to be altogether unaffected.