Abstract I consider recent observations of ultramassive black holes. These appear to confirm theoretical predictions that the relation between central black hole mass M and spheroid velocity dispersion σ has the same form M∝σ4 in spiral galaxies, elliptical galaxies, and cluster ellipticals, but has differing normalizations. These arise from the need for longer black hole accretion episodes to expel the gas otherwise potentially able to feed the holes in the latter two types of host. In a sample drawn from a mixture of galaxy host types the fitted power of σ will slightly exceed the theoretically–derived value of 4 because of the differing normalizations. The observed hole masses do not currently reach the theoretical maximum values possible for disc accretion, set by the equality of the ISCO and self–gravity radii, probably because the host galaxies have insufficient gas.
Andrew King (Sat,) studied this question.