It is important to note that BH mass does not correlate with disks in the same way that it does with bulges. Figure 4 shows the correlations of BH mass with (left) bulge and (right) total luminosity. Figure 4 (right) shows that disk galaxies with small bulge-to-total luminosity ratios destroy the reasonably good correlation seen in Figure 4 (left). In addition, Figure 4 shows four galaxies that have strong BH mass limits but no bulges. They further emphasize the conclusion that disks do not contain BHs with nearly the same mass fraction as do bulges. In particular, in the bulgeless galaxy M33, the upper limit on a BH mass from STIS spectroscopy is M 1000 M. If M33 contained a BH with the median mass fraction observed for bulges, then we would expect that M ~ 3 × 107 M.
Figure 4 tells us that BH masses do not "know about" galaxy disks. Rather, they correlate with the high-density bulge-like component in galaxies.
Figure 4. (left) M - MB, bulge correlation from Figure 1. (right) Plot of M against the total absolute magnitude of the host galaxy. Filled symbols denote elliptical galaxies, open symbols denote bulges of disk galaxies. Crosses denote galaxies that do not contain a bulge: M33 is from Gebhardt et al. (2001); IC 342 is from Böker et al. (1999), and NGC 4395 is from Filippenko & Ho (2001).
These results do not preclude BHs in pure disk galaxies as long as they are small. Filippenko & Ho (2001) emphasize that some pure disks are Seyfert galaxies. They probably contain BHs. An extreme example is NGC 4395, the lowest-luminosity Seyfert known (Fig. 4). However, if its BH were radiating at the Eddington rate, then its mass would be only M ~ 100 M (Filippenko & Ho 2001). So disks can contain BHs, but their masses are much smaller in relation to their disk luminosities than are bulge BHs in relation to bulge luminosities. It is possible that the small BHs in disks are similar to the seed BHs that once must have existed even in protobulges before they grew monstrous during the AGN era.