Although incontrovertible observational proof of the existence of supermassive black holes (SMBHs) has yet not been found, evidence is mounting to suggest the presence of massive dark objects, or large mass concentrations at the centres of galaxies. Black holes, by definition, cannot be `seen' and instead one must look for the consequences of their presence. The presence of SMBHs has been inferred indirectly from the energetics of accretion required to power luminous AGN and explain rapid flux variability and, more directly, from kinematic studies of the influence of the black hole's gravitational pull on stars and gas orbiting close to it in the central regions of both active and non-active galaxies. Theoretical models rule out alternatives to a supermassive black holes such as collections of brown or white dwarf stars, neutron stars or stellar-mass black holes which would merge and shine or evaporate too quickly (Maoz 1995, 1998; Genzel et al. 1997, 2000).
4.1. Quasar lifetimes and the black hole legacy
Soon after the discovery of quasars it became clear that they were most common when the Universe was relatively young with the peak of the quasar epoch at redshift z ~ 2.5 or a look-back time of 65% of the age of the Universe (See Figure 4); today bright quasars are rare and weaker Seyferts dominate instead. The number of dead quasars or relic, dormant black holes left today can estimated by applying some simple arguments to the quasar observations. Soltan (1982) integrated the observed light emitted by quasars, and, assuming the power source for quasar light is accretion of material by a supermassive black hole with a mass-to-energy conversion efficiency of 10% and that the black hole grows during the active phase, predicted the total mass in relic black holes today. Knowing the number of galaxies per unit volume of space (e.g. Loveday et al. 1992), if one assumes that all galaxies went through a quasar phase at some time in their lives, then each galaxy should, on average, contain a ~ 108 M black hole as a legacy of this violent, but short-lived period (~ 107 to ~ 108 years). Alternatively, if only a small fraction of galaxies went through a quasar phase, the active phase would have lasted lasted longer (> 109 years) and the remnant SMBHs would be relatively rare, but unacceptably massive (> 109 M) (e.g. Cavaliere et al., 1983; Cavaliere & Szalay 1986, Cavaliere & Padovani 1988).
Figure 4. Relative number of galaxies per unit volume of space (as a fraction of the peak value) detected in the SLOAN Digital Sky Survey, as a function of look-back time i.e. time running backwards from now to the Big Bang [using data from Schneider et al. (2002)].
More complex models including quasar evolution (e.g. Tremaine 1996; Faber et al. 1997) and the effects of galaxy growth (e.g. Haehnelt & Rees 1993) favour short-lived periods of activity in many generations of quasars, or a mixture of continuous and recurrent activity (Small & Blandford 1992; Cen 2000; Choi, Yang & Yi 2001). The complex physics of accretion and black hole growth, however, remain an area of active study (e.g. Blandford & Begelman 1999; Fabian 1999). Nevertheless, the range of black hole mass of interest is thought to be M ~ 106 to 109.5 M, with the lower mass holes being ubiquitous (Kormendy & Gebhardt 2001).
4.2. Irresistible black holes - dynamics of gas and stars
Although the prodigious energy outputs from powerful quasars offer strong circumstantial evidence that supermassive black holes exist, most notably in driving the ejection and acceleration of long, powerful jets of plasma close to the speed of light (Rees et al. 1982), it has not, until recently, been possible to make more direct kinematic measurements of the black hole's gravitational influence. The mass of a central object, the circular velocity of an orbiting star and the radius of the orbit are related by Newton's Laws of motion and gravity. Precise measurements of the velocities of stars and gas close to the centre of a galaxy are then used to determine the mass of the central object.
The strongest dynamical evidence for black holes comes from studies of centre of our own Galaxy and a nearby Seyfert, NGC 4258; a decade of painstaking observations of a cluster of stars orbiting around the mildly active centre of the Milky Way, within a radius of 0.07 light years of the central radio source Sgr A*, suggest a central mass of M = (2.6 ± 0.2) × 106 M (Eckart & Genzel 1997; Genzel et al. 1997, 2000; Ghez 2000). Discovery of strong radio spectral lines, or megamasers, emitted from water molecules in a rapidly rotating nuclear gas disc at the centre of NGC 4258 implies a centre mass M = (4 ± 0.1) × 107 M concentrated in a region smaller than 0.7 light years (Miyoshi et al. 1995), again small enough to rule out anything other than a black hole (Maoz 1995, 1998). Precision measurements of black hole masses in other galaxies using a variety of techniques, although challenging and still model dependent, have become increasingly common (e.g. Maggorian et al. 1998; Bower et al. 1998; Gebhardt et al. 2000) and now more than 60 active and non-active galaxies have black hole estimates.