4 Discussion and conclusions

In Bertola et al. (1998) we demonstrated that it is possible to detect the signature of a CNKD in the emission-line PV diagrams obtained from ground-based spectroscopy of nearby galaxies. In addition, we showed that once properly modelled, this technique can lead to reliable upper-limits for the central mass concentration. Using these results, and to identify new galaxies hosting a CNKD, we obtained long-slit spectra along the major axes of 23 nearby disk galaxies, measuring the PV diagrams of the ionized-gas components from the emission lines.

To achieve our goal, we analyzed these emission-line PV diagrams by measuring the values of the velocity gradient, the velocity dispersion, and the integrated flux of the ionized gas at different radii, and by comparing the shape of the observed emission lines with that predicted for a thin disk of gas moving in circular orbits in the galaxy plane. This classification allowed us to recognize the possible presence of a CNKD in 3 of our objects. Recently, Rubin et al. (1997) and Sofue et al. (1998) discussed the kinematics of rapidly-rotating gas disks observed in the central few hundred parsecs of S0's and spiral galaxies. By combining our sample with their samples, we find that the PV diagrams of 9 out 54 galaxies exhibit a $\Gamma>2$, which is indicative of a CNKD. The fact that in the majority of these cases CNKDs have not been observed means that not all of these galaxies enough gas detectable and rotating Keplerian orbits. Therefore, we estimated that the frequency of CNKDs, measured from ground-based spectroscopy of emission-line disk galaxies, is $\la$20%. This result is consistent with the findings of Sarzi et al. (2001), which are based on HST spectroscopy. Indeed, they found a gaseous disk with a well-ordered velocity field suitable for dynamical modelling at HST resolution in only 4 of the 23 randomly selected emission-line disk galaxies they observed.

The analysis of ground-based PV diagrams allows identification of those galaxies that are good candidates for hosting a CNKD rotating around a central mass concentration, and are therefore are good candidates for follow-up HST spectroscopy. It is worth noting that in this way it is possible to improve the present low success rate of HST programs aimed at estimating SMBH masses in nearby bulges by modelling nuclear gas kinematics. Although the SMBH mass hosted by the candidates selected by ground-based observations are expected to lie in the high-mass end of the $M_\bullet-\sigma$ diagram because of the resolution limits imposed on their PV diagrams by seeing, these candidates are required to elucidate the relationship for disk galaxies, which are underrepresented relative to elliptical galaxies in the sample of galaxies studied so far.

Acknowledgements

We thank Betsy Gillespie for reading the manuscript. This research has made use of the Lyon-Meudon Extragalactic Database (LEDA) and Digitized Sky Survey (DSS).