The recent completion of an optical spectroscopic survey of the nuclei of a large sample of nearby galaxies (Ho, Filippenko, & Sargent 1995) provides an excellent data base with which to reexamine the issues discussed above. High signal-to-noise ratio, moderate-resolution (2.5-4 Å), long-slit spectra were obtained for a magnitude-limited (BT 12.5 mag) sample of 486 northern ( > 0°) galaxies using the Hale 5-m reflector at Palomar Observatory (Filippenko & Sargent 1985). Nearly all of the observations were obtained with a 2" slit, and spectra were derived from an effective aperture of 2" x 4", corresponding to physical dimensions of ~ 170 pc x 340 pc for the median distance of 17 Mpc.
Approximately 60% of S0-Sm spirals in our sample contain bars (Fig. 1), in agreement with the statistics of Sellwood & Wilkinson (1993) for field spirals. Adopting the spectral classifications given in Ho, Filippenko, & Sargent (1996b), we examine the dependence of the detection rate of H II nuclei (those photoionized by OB stars) and AGNs [including LINERs (low-ionization nuclear emission-line regions), LINER/H II nuclei transition objects, and Seyfert 1 and 2 nuclei] on the presence of a bar for 428 spiral galaxies. While the distribution of unbarred galaxies does not differ appreciably from that of barred galaxies among AGNs, there is clearly a difference among H II nuclei (Fig. 2). Applying the Kolmogorov-Smirnov test, we find that the probability (PKS) that the two distributions are drawn from the same population is only 50.1% for the AGNs, whereas PKS = 0.2% for the H II nuclei. The latter result is highly statistically significant, implying that nuclear star formation occurs more frequently in barred spirals.
Figure 1. Distribution of Hubble types for the 486 galaxies in the survey. Unbarred galaxies are shown in the hatched histogram, and barred galaxies in the unhatched histogram.
Figure 2. Distribution of AGNs (top) and H II nuclei (bottom) as a function of Hubble type. Barred and unbarred galaxies are shown separately.
The effect of bars on the strength of the activity can be gauged by inspecting the distributions of the luminosity of a strong emission line such as H. In Figure 3, ``early-type'' spirals (S0-Sbc) are shown separately from ``late-type'' spirals (Sc-Sm). It is apparent that among H II nuclei, barred early-type systems show significantly higher H luminosities than unbarred counterparts (PKS = 3.3%), whereas no such trend is present for the late-type systems. The line strengths of AGNs, by contrast, appear not to be influenced by the presence of a bar, regardless of the morphological type. We find the same trend when we use the equivalent width of H emission as the indicator instead of the line luminosity (Ho et al. 1996c).
Figure 3. Distributions of H luminosities for (a) H II nuclei and (b) AGNs. The two top panels in each case show barred and unbarred for early-type (S0-Sbc) spirals, while the two bottom panels show late-type (Sc-Sm) spirals.